Post on 09-May-2020
transcript
Heritage Language for Chinese Australians: Negotiating ‘Chineseness’ and
Capitalising on Resources in the Lived World
Dr. Guanglun (Michael) Mu 穆光伦
Bachelor of Science (Beijing Normal University)
Master of Educational Administration (The University of New South Wales)
Principal Supervisor: Associate Professor Dr. Karen Dooley
Associate Supervisor: Dr. Paul Shield
Associate Supervisor: Dr. Catherine Doherty
Faculty of Education
Queensland University of Technology
Submitted in fulfilment of the requirements for the degree of Doctor of Philosophy
2013
i
Keywords
Heritage Language, Chinese Heritage Language proficiency, Heritage Language Learner,
Chinese Australians, motivation, investment, capital, ethnic identity, Chineseness, habitus,
field, Bourdieu, mixed methods design, Structural Equation Modelling
ii
Abstract
The ethnic identity and commitment of Heritage Language Learners play salient roles in
Heritage Language learning process. The mutually constitutive effect amongst Heritage
Language Learner’s ethnic identity, commitment, and Heritage Language proficiency has
been well documented in social psychological and poststructuralist literatures. Both social
psychological and poststructural schools offer meaningful insights into particular contexts but
receive critiques from other contexts. In addition, the two schools largely oppose each other.
This study uses Bourdieu’s sociological triad of habitus, capital, and field to reconcile the two
schools through the examination of Chinese Heritage Language Learners in Australia, an
idiosyncratic social, cultural, and historical context for these learners. Specifically, this study
investigates how young Chinese Australian adults (18-35 in age) negotiate their ‘Chineseness’
and capitalise on resources through Chinese Heritage Language learning in the lived world.
The study adopts an explanatory mixed methods design to combine the quantitative
approach with the qualitative approach. The initial quantitative phase addresses the first
research question: Is Chinese Heritage Language proficiency of young Chinese Australian
adults influenced by their investment of capital, the strength of their habitus of ‘Chineseness’,
or both? The subsequent qualitative phase addresses the second research question: How do
young Chinese Australian adults understand their Chinese Heritage Language learning in
relation to (potential) profits produced by this linguistic capital in given fields?
The initial quantitative phase applies Structural Equation Modelling to analyse the data
from an online survey with 230 respondents. Findings indicate the statistically significant
positive contribution made by the habitus of ‘Chineseness’ and by investment of capital to
Chinese Heritage Language proficiency (r = .71 and r = .86 respectively). Subsequent
multiple regression analysis demonstrates that 62% of the variance of Chinese Heritage
iii
Language proficiency can be accounted for by the joint contribution of ‘Chineseness’ and
‘capital’. The qualitative phase of the study uses multiple interviews with five participants. It
reveals that Chinese Heritage Language offers meaningful benefits for participants in the
forms of capital production and habitus capture or recapture. Findings from the two phases
talk to each other in terms of the inherent entanglement amongst habitus of ‘Chineseness’,
investment of capital, and Chinese Heritage Language proficiency.
The study offers important contributions. Theoretically, by virtue of Bourdieu’s
signature concepts of habitus, capital, and field, the study provides answers to questions that
both social psychological and poststructuralist theories have long been struggling to answer.
Methodologically, the position of ‘pluralism’ talks back to Bourdieu’s theory and forwards to
the mixed methods design. Particularly, the study makes a methodological breakthrough: A
set of instruments was developed and validated to quantify Bourdieu’s key concepts of
capital and habitus within certain social fields. Practically, understanding Chinese Australians’
heterogeneity and the potential drivers behind Chinese Heritage Language learning
contributes to the growing interest in Chinese Australians’ contemporary life experiences and
helps to better accommodate linguistically diverse Chinese Heritage Language Learners in
Chinese language courses. In addition, this study is very timely. It resonates with the recently
released Australia in the Asian Century White Paper: Chinese Australians, with sound
knowledge of Chinese culture and language obtained through negotiating their ‘Chineseness’
and capitalising on diverse resources for learning, will help to serve Australia’s economic,
social, and political needs in unique ways.
iv
Table of Contents
Keywords ................................................................................................................................................................i
Abstract ................................................................................................................................................................. ii
Table of Contents .................................................................................................................................................iv
List of Figures ..................................................................................................................................................... vii
List of Tables ...................................................................................................................................................... viii
List of Abbreviations ............................................................................................................................................. x
Statement of Original Authorship .................................................................................................................... xii
Acknowledgement ............................................................................................................................................. xiii
CHAPTER 1: INTRODUCTION ...........................................................................................1
1.1 Definitions ........................................................................................................................................................ 5 1.1.1 Definitions of Heritage Language .............................................................................................................. 6 1.1.2 Heritage Language Learners ...................................................................................................................... 6 1.1.3 Heritage Language Learners and Non-Heritage Language Learners ......................................................... 9
1.2 Background ................................................................................................................................................... 11 1.2.1 Cultural and language policies in Australia ............................................................................................. 12 1.2.2 Chinese immigrants in Australia .............................................................................................................. 20 1.2.3 Chinese language in Australia .................................................................................................................. 26
1.3 Researcher’s subjectivity .............................................................................................................................. 28
1.4 Significance of the research .......................................................................................................................... 29
1.5 Thesis structure ............................................................................................................................................. 30
CHAPTER 2: LITERATURE REVIEW .............................................................................33
2.1 Commitment to CHL learning: ‘Motivation’ and ‘investment’ ................................................................ 33 2.1.1 Motivation: A social psychological perspective ....................................................................................... 34 2.1.2 Investment: A poststructuralist perspective .............................................................................................. 39
2.2 Identity construction in CHL learning ........................................................................................................ 44 2.2.1 Identity: A social psychological perspective ............................................................................................ 44 2.2.2 Identity: A poststructuralist perspective ................................................................................................... 48
2.3 Reframing through a Bourdieusian stance ................................................................................................. 54 2.3.1 ‘Motivation’, ‘investment’, and ‘capital’ ................................................................................................. 56 2.3.2 ‘Ethnic identity’ and ‘habitus’ .................................................................................................................. 60
2.4 Chapter summary ......................................................................................................................................... 62
CHAPTER 3: THEORETICAL FRAMEWORK ...............................................................65
3.1 Capital ............................................................................................................................................................ 66
v
3.1.1 Economic capital ..................................................................................................................................... 67 3.1.2 Cultural capital ........................................................................................................................................ 67 3.1.3 Social capital ........................................................................................................................................... 69 3.1.4 Symbolic capital ...................................................................................................................................... 70 3.1.5 Capitalising on language: Linguistic capital ........................................................................................... 72 3.1.6 Section summary ..................................................................................................................................... 75
3.2 Field ............................................................................................................................................................... 76
3.3 Habitus .......................................................................................................................................................... 81 3.3.1 Conceptualising habitus .......................................................................................................................... 82 3.3.2 Interpreting ‘Chineseness’ through habitus ............................................................................................. 89 3.3.3 Confucian dispositions ............................................................................................................................ 94
3.4 Chapter summary ...................................................................................................................................... 103
CHAPTER 4: METHODOLOGY ..................................................................................... 107
4.1 Philosophical position ................................................................................................................................ 107
4.2 Mixed methods research ............................................................................................................................. 110
4.3 Quantitative phase....................................................................................................................................... 114 4.3.1 Research questions and hypotheses ........................................................................................................ 115 4.3.2 Target population .................................................................................................................................... 116 4.3.3 Instrument design and development ....................................................................................................... 118 4.3.4 Pilot study ............................................................................................................................................. 126 4.3.5 Main study ............................................................................................................................................ 128
4.4 Qualitative phase ........................................................................................................................................ 139 4.4.1 Thematising ........................................................................................................................................... 141 4.4.2 Designing .............................................................................................................................................. 142 4.4.3 Interviewing .......................................................................................................................................... 149 4.4.4 Transcribing .......................................................................................................................................... 153 4.4.5 Analysing .............................................................................................................................................. 154 4.4.6 Reporting ............................................................................................................................................... 156
4.5 Ethical considerations ................................................................................................................................ 157
4.6 Chapter summary ...................................................................................................................................... 160
CHAPTER 5: QUANTITATIVE DATA ANALYSIS AND FINDINGS ......................... 161
5.1 Pilot study ................................................................................................................................................... 161
5.2 Main study .................................................................................................................................................. 168 5.2.1 Validity and reliability of measurement models: EFA and CFA ............................................................ 172 5.2.2 Hypothesis testing: SEM ....................................................................................................................... 223 5.2.3 Habitus and capital working together: Regression ................................................................................ 234
5.3 Chapter summary ...................................................................................................................................... 241
CHAPTER 6: QUALITATIVE DATA ANALYSIS AND FINDINGS ............................ 243
6.1 How were the five participants selected? ................................................................................................. 243
6.2 Who were the five participants? ................................................................................................................ 246
vi
Adam: “I just need to catch up!” ..................................................................................................................... 246 Bob: “Chinese is part of me.” ......................................................................................................................... 248 Crystal: Chinese entertainment stuff is “really funny”. .................................................................................. 249 Dianna: “I felt how important (it was) to speak my own home language!” .................................................... 249 En-ning: “Learning Chinese is definitely something I am going to pursue in the rest of my life.” ................ 250
6.3 What did CHL mean to the five participants? .......................................................................................... 251 6.3.1 Potential production of economic capital through CHL learning ........................................................... 252 6.3.2 Production of cultural capital through CHL learning ............................................................................. 253 6.3.3 Production of social capital through CHL learning ................................................................................ 257 6.3.4 Production of symbolic capital through CHL learning .......................................................................... 261 6.3.5 Making sense of the habitus of ‘Chineseness’ through CHL learning ................................................... 263 6.3.6 Capital in different fields ....................................................................................................................... 268 6.3.7 Catering to the field of forces ................................................................................................................ 273
6.4 Are the findings consistent? ........................................................................................................................ 276 6.4.1 Capital and CHL proficiency ................................................................................................................. 276 6.4.2 ‘Chineseness’ and CHL proficiency ....................................................................................................... 280 6.4.3 Quantitative coda ................................................................................................................................... 283
6.5 Chapter summary ....................................................................................................................................... 287
CHAPTER 7: DISCUSSION AND CONCLUSION .........................................................291
7.1 Discussions ................................................................................................................................................... 291 7.1.1 Habitus, capital, and practice in fields ................................................................................................... 292 7.1.2 Field of forces ........................................................................................................................................ 297
7.2 Contributions of the study .......................................................................................................................... 300 7.3.1 Originality and contributions to new knowledge ................................................................................... 301 7.3.2 Methodological contributions ................................................................................................................ 302 7.3.3 Practical contributions ........................................................................................................................... 303
7.3 Limitations of the study .............................................................................................................................. 306
7.4 Future directions ......................................................................................................................................... 308
7.5 Overall conclusions of the study ................................................................................................................ 309
Appendices ......................................................................................................................................................... 315 Appendix 1: Information sheet for participation in the online survey ............................................................ 315 Appendix 2: Information sheet for participation in the interview ................................................................... 316 Appendix 3: The consent form ....................................................................................................................... 317 Appendix 4: Interview schedule ..................................................................................................................... 318
Bibliography ...................................................................................................................................................... 321
vii
List of Figures
Figure 2.1. Summary of literature review .................................................................................................... 63 Figure 3.1. Forms of capital and language ................................................................................................... 75 Figure 3.2. Forms of capital and language in fields ..................................................................................... 80 Figure 3.3. Habitus and language in fields ................................................................................................... 89 Figure 3.4. Theoretical framework ............................................................................................................. 105 Figure 4.1. Theoretical framework guiding the quantitative phase ............................................................ 116 Figure 4.2. Structural relationship between ‘Chineseness’ and ‘CHL proficiency’ ................................... 138 Figure 4.3. Structural relationship between ‘capital’ and ‘CHL proficiency’ ............................................ 138 Figure 4.4. Theoretical framework guiding the qualitative phase .............................................................. 142 Figure 5.1. Age distribution of participants ................................................................................................ 169 Figure 5.2. Distribution of participants’ resident cities .............................................................................. 169 Figure 5.3. Composition of participants’ birthplace ................................................................................... 170 Figure 5.4. Composition of participants’ generation .................................................................................. 170 Figure 5.5. Composition of participant’s language usage at home ............................................................. 171 Figure 5.6. Composition of Chinese dialects used at home ....................................................................... 171 Figure 5.7. Scree plot of the ‘Chineseness’ scale ....................................................................................... 177 Figure 5.8. Measurement model for the construct ‘Chineseness’ .............................................................. 178 Figure 5.9. Modified measurement model for the construct ‘Chineseness’ ............................................... 181 Figure 5.10. Scree plot of the ‘economic capital’ scale .............................................................................. 187 Figure 5.11. Measurement model for the construct ‘economic capital’ ..................................................... 188 Figure 5.12. Scree plot of the ‘cultural capital’ scale ................................................................................. 193 Figure 5.13. Measurement model for the construct ‘cultural capital’ ........................................................ 194 Figure 5.14. Modified measurement model for the construct ‘cultural capital’ ......................................... 196 Figure 5.15. Scree plot of the ‘social capital’ scale .................................................................................... 201 Figure 5.16. Measurement model for the construct ‘social capital’ ........................................................... 202 Figure 5.17. Scree plot of the ‘symbolic capital’ scale ............................................................................... 206 Figure 5.18. Measurement model for the construct ‘symbolic capital’ ...................................................... 207 Figure 5.19. Modified measurement model for the construct ‘symbolic capital’ ...................................... 210 Figure 5.20. Scree plot of the ‘CHL proficiency’ scale .............................................................................. 215 Figure 5.21. Measurement model for the construct ‘CHL proficiency’ ..................................................... 216 Figure 5.22. Modified measurement model for the construct ‘CHL proficiency’ ...................................... 219 Figure 5.23. Structural model for the relationship between ‘Chineseness’ and ‘CHL proficiency’ ........... 224 Figure 5.24. Scree plot of the ‘capital’ scale .............................................................................................. 229 Figure 5.25. Measurement model for the construct ‘capital’ ..................................................................... 230 Figure 5.26. Structural model for the relationship between ‘capital’ and ‘CHL proficiency’ .................... 232 Figure 5.27. Histogram and P-P plots of the standardised residuals of the ‘CHL proficiency’.................. 239 Figure 5.28. Scatter plots of the standardised and the studentised residuals of ‘CHL proficiency’ against the
standardised predicted values of ‘CHL proficiency’ .................................................................................. 240 Figure 5.29. Partial plots of the residuals of ‘CHL proficiency’ against ‘capital’ and ‘Chineseness’ ........ 240
viii
List of Tables
Table 5.1 Item sets and their corresponding constructs in the pilot study .................................................. 164 Table 5.2 Summary results of internal consistency reliability test ............................................................. 165 Table 5.3 Item sets and their corresponding constructs in the main study ................................................. 166 Table 5.4 Descriptive statistics of indicators .............................................................................................. 172 Table 5.5 Correlation matrix for the indicators within the construct ‘Chineseness’ ................................... 174 Table 5.6 KMO and Bartlett's test for the indicators within the construct ‘Chineseness’ .......................... 175 Table 5.7 Communalities for the indicators within the construct ‘Chineseness’ ........................................ 176 Table 5.8 Total variance of the construct ‘Chineseness’ explained by its indicators .................................. 176 Table 5.9 Assessment of normality for the indicators within the construct ‘Chineseness’ ......................... 177 Table 5.10 Model fit indices for the construct ‘Chineseness’ ..................................................................... 178 Table 5.11 Standardised residual covariances for the indicators within the construct ‘Chineseness’ ........ 179 Table 5.12 Modification indices for the construct ‘Chineseness’ ............................................................... 180 Table 5.13 Assessment of normality for the indicators within the construct ‘Chineseness’ (modified) ..... 180 Table 5.14 Model fit indices for the construct ‘Chineseness’ (modified) ................................................... 181 Table 5.15 Proportional factor score weights for the indicators within the construct ‘Chineseness’ ......... 182 Table 5.16 Correlation matrix for the indicators within the construct ‘economic capital’ ......................... 185 Table 5.17 KMO and Bartlett's test for the indicators within the construct ‘economic capital’ ................. 185 Table 5.18 Communalities for the indicators within the construct ‘economic capital’ .............................. 186 Table 5.19 Total variance of the construct ‘economic capital’ explained by its indicators ......................... 186 Table 5.20 Assessment of normality for the indicators within the construct ‘economic capital’ ............... 187 Table 5.21 Model fit indices for the construct ‘economic capital’ ............................................................. 189 Table 5.22 Proportional factor score weights for the indicators within the construct ‘economic capital’ .. 189 Table 5.23 Correlation matrix for the indicators within the construct ‘cultural capital’............................. 191 Table 5.24 KMO and Bartlett's test for the indicators within the construct ‘cultural capital’ .................... 191 Table 5.25 Communalities for the indicators within the construct ‘cultural capital’ .................................. 192 Table 5.26 Total variance of the construct ‘cultural capital’ explained by its indicators ............................ 192 Table 5.27 Assessment of normality for the indicators within the construct ‘cultural capital’ ................... 193 Table 5.28 Model fit indices for the construct ‘cultural capital’ ................................................................. 194 Table 5.29 Standardised residual covariances for the indicators within the construct ‘cultural capital’ .... 195 Table 5.30 Modification indices for the construct ‘cultural capital’ ........................................................... 195 Table 5.31 Assessment of normality for the indicators within the construct ‘cultural capital’ (modified) . 195 Table 5.32 Model fit indices for the construct ‘cultural capital’ (modified) ............................................... 197 Table 5.33 Proportional factor score weights for the indicators within the construct ‘cultural capital’ ..... 197 Table 5.34 Correlation matrix for the indicators within the construct ‘social capital’ ............................... 199 Table 5.35 KMO and Bartlett's test for the indicators within the construct ‘social capital’ ....................... 199 Table 5.36 Communalities for the indicators within the construct ‘social capital’ ..................................... 200 Table 5.37 Total variance of the construct ‘social capital’ explained by its indicators ............................... 200 Table 5.38 Assessment of normality for the indicators within the construct ‘social capital’...................... 201 Table 5.39 Model fit indices for the construct ‘social capital’ ................................................................... 203 Table 5.40 Proportional factor score weights for the indicators within the construct ‘social capital’ ........ 203 Table 5.41 Correlation matrix for the indicators within the construct ‘symbolic capital’ .......................... 205 Table 5.42 KMO and Bartlett's test for the indicators within the construct ‘symbolic capital’ .................. 205 Table 5.43 Communalities for the indicators within the construct ‘symbolic capital’ ............................... 206 Table 5.44 Total variance of the construct ‘symbolic capital’ explained by its indicators.......................... 206 Table 5.45 Assessment of normality for the indicators within the construct ‘symbolic capital’ ................ 207 Table 5.46 Model fit indices for the construct ‘symbolic capital’ .............................................................. 208 Table 5.47 Standardised residual covariances for the indicators within the construct ‘symbolic capital’ .. 209 Table 5.48 Modification indices for the construct ‘symbolic capital’ ........................................................ 209 Table 5.49 Assessment of normality for the indicators within the construct ‘symbolic capital’ (modified)
.................................................................................................................................................................... 209 Table 5.50 Model fit indices for the indicators within the construct ‘symbolic capital’ ............................ 211 Table 5.51 Proportional factor score weights for the indicators within the construct ‘symbolic capital’ ... 211 Table 5.52 Correlation matrix for the indicators within the construct ‘CHL proficiency’ ......................... 213 Table 5.53 KMO and Bartlett's test for the indicators within the construct ‘CHL proficiency’ ................. 214
ix
Table 5.54 Communalities for the indicators within the construct ‘CHL proficiency’ .............................. 214 Table 5.55 Total variance of the construct ‘CHL proficiency’ explained by its indicators ......................... 215 Table 5.56 Assessment of normality for the indicators within the construct ‘CHL proficiency’ ............... 216 Table 5.57 Model fit indices for the construct ‘CHL proficiency’ ............................................................. 217 Table 5.58 Standardised residual covariances for the indicators within the construct ‘CHL proficiency’ . 218 Table 5.59 Modification indices for the construct ‘CHL proficiency’ ....................................................... 218 Table 5.60 Assessment of normality for the indicators within the construct ‘CHL proficiency’ (modified)
.................................................................................................................................................................... 219 Table 5.61 Model fit indices for the construct ‘CHL proficiency’ (modified) ........................................... 220 Table 5.62 Proportional factor score weights for the indicators within the construct ‘CHL proficiency’ .. 220 Table 5.63 Item sets and their corresponding constructs in the (modified) measurement models ............. 222 Table 5.64 Model fit indices for the structural model: ‘Chineseness’ and ‘CHL proficiency’ ................... 225 Table 5.65 Formulae for computation of the four composite variables ...................................................... 226 Table 5.66 Correlation matrix for the indicators within the construct ‘capital’ ......................................... 227 Table 5.67 KMO and Bartlett's test for the indicators within the construct ‘capital’ ................................. 228 Table 5.68 Communalities for the indicators within the construct ‘capital’ ............................................... 228 Table 5.69 Total variance of the construct ‘capital’ explained by its indicators ......................................... 228 Table 5.70 Assessment of normality for the indicators within the construct ‘capital’................................ 229 Table 5.71 Model fit indices for the construct ‘capital’ ............................................................................. 230 Table 5.72 Proportional factor score weights for the indicators within the construct ‘capital’ .................. 231 Table 5.73 Model fit indices for the structural model: ‘Capital’ and ‘CHL proficiency’ ........................... 233 Table 5.74 Formulae for computation of the constructs ‘Chineseness’, ‘capital’, and ‘CHL proficiency’ 235 Table 5.75 Correlation matrix for the variables .......................................................................................... 235 Table 5.76 Model summary ........................................................................................................................ 237 Table 5.77 ANOVA results ......................................................................................................................... 237 Table 5.78 Model parameters ..................................................................................................................... 237 Table 5.79 Collinearity diagnostics ............................................................................................................ 238 Table 6.1 Case-wise diagnostics ................................................................................................................. 244 Table 6.2 Overview of participants ............................................................................................................ 246 Table 6.3 Correlation between forms of capital and social practices ......................................................... 284 Table 6.4 Interview data summary ............................................................................................................. 288 Table 6.5 Contributing factors to CHL proficiency .................................................................................... 289
x
List of Abbreviations
ABS Australian Bureau of Statistics
ACARA Australian Curriculum Assessment and Reporting Authority
ANOVA Analysis of Variance
CFA Confirmatory Factor Analysis
CFI Comparative Fit Index
CHL Chinese Heritage Language
CHLL(s) Chinese Heritage Language Learner(s)
COAG Council of Australian Government
DEET Department of Employment, Education, and Training
EFA Exploratory Factor Analysis
ESL English as a Second Language
H0 Null Hypothesis
HL Heritage Language
HLL(s) Heritage Language Learner(s)
IFI Incremental Fit Index
KMO Kaiser-Meyer-Olkin
NFI Normed Fit Index
NHLL(s) Non-Heritage Language Learner(s)
RFI Relative Fit Index
RMSEA Root Mean Square Error of Approximation
RQ (s) Research Question(s)
SEM Structural Equation Modelling
SMC Squared Multiple Correlations
xii
Statement of Original Authorship
The work contained in this thesis has not been previously submitted to meet requirements for
an award at this or any other higher education institution. To the best of my knowledge and
belief, the thesis contains no material previously published or written by another person
except where due reference is made.
Signature:
Date: 17 June 2013
xiii
Acknowledgement
My PhD project was completed under the sponsorship of Queensland University of
Technology (QUT). I would like to take this opportunity to acknowledge the substantial
financial assistance that I received from the QUT Postgraduate Research Award.
I have complex and rich feelings about this PhD journey. During the entire life of this
journey, there have been struggles and passions, frustrations and achievements, stress and
pleasure, suffering and benefits, loneliness and teamwork, as well as many other feelings.
However, I was not alone along this journey. I felt how lucky I was to have so many people
around me, supervising me, helping me, supporting me, and encouraging me. I am deeply
grateful to these people.
First and foremost, I would like to extend my utmost gratitude to my supervisory team,
Associate Prof. Karen Dooley, Dr. Catherine Doherty, and Dr. Paul Shield, who introduced
me to this body of scholarship, guided me through this project, and pushed me to be the best
that I can. They were so engaged in my project that they contributed significant time to
supervise me, without which I would have never ever come to the end of this PhD journey. I
will remember their inculcation all through my life. As the Chinese saying goes, “一日为师,
终身为师” (Be my teacher for a day, be my teacher for a lifetime). Special thanks are also
given to Prof. Allan Luke, who shines the lights to guide me down the academic road that I
am walking along.
Secondly, my participants volunteered their time to complete the online questionnaire.
Five of them spent extra time to participate in the subsequent interviews. They shared their
stories and life experiences with me and helped me produce and enrich my data. I have given
my thanks to the five participants before and after the interviews, but I have no way to thank
the rest of my participants in person because they did not leave their contact details. I
xiv
therefore owe thanks to these people. This feeling urged me to work even harder on my
project. I have tried my best to improve the quality of this thesis to respect the contribution of
their time.
Thirdly, I would like to express my sincere gratitude both to the internal reviewers at
the confirmation and the final stage, and to the external examiners of my thesis. Their
constructive comments and thought-provoking advice helped me bring my thesis to a higher
level. I am also thankful to the staff from the former Centre for Learning Innovation for their
‘scaffolding’ role along my PhD journey: Prof. Carmel Diezmann, Ms. Jeannean Botha, Ms.
Jennifer Yared, Dr. Briony Wainman, Dr. Mary Clowes, Prof. Jo Brownlee, Associate Prof.
Julie Davis, Dr. Jo Lampert, Ms. Carol Partridge, Ms. Helen Tanner, and Ms. Rowena
McGregor.
Next, I would like to extend my thanks, from the depth of my heart, to all my friends
who supported me during the entire three years of my PhD journey, both in academic and in
daily life: Dr. Yaxing Zhang, Dr. Yang Hu, Ms. Jianwei Liu, Mr. Xiangguo Wang, Dr. Jeanne
Yang, Ms. Xinxin Fan, Dr. Juncheng Dai, Ms. Qiuxiang Huang, Dr. Janet Zhong, Dr.
Katherine Hanna, Mr. Charles Allen, Mr. Ark Du, Mr. Rui Jiang, Ms. Tracy Liu, Mr. Saif
Alamri, Dr. Jill Fox, Dr. Yuming Guo, Ms. Shandy Li, Dr. Ingrid Wang, Ms. Pei Liu, Dr.
Dagang Wang, Dr. Janet Hou, Associate Prof. Sue Walker, Dr. Louise Mercer, Dr. Lyn May,
Prof. Kar-Tin Lee, Mr. Christopher Meakin, Emeritus Prof. Sandra Vianne McLean A.M., Dr.
Debby Syu, Mr. Leon Zhang, Dr. Xiaofang Ye, Mr. Juming Shen, Ms. Maggie Kim, Dr.
Susan Sim, Dr. Li Yuan, Associate Prof. Lisa Ehrich, and Associate Prof. Deborah Henderson.
Without the friendship of these people, my PhD journey would have been a lonely one.
Last but not least, I am indebted to my beloved family members: Ms. Weiming Li (李
伟明), Mr. Shuhuai Mu (穆书淮), and Mr. Shizhuo Gui (桂士卓), who accompanied me
xv
through the ups and downs along my PhD journey, helped me build up my confidence, and
taught me never to give up. I hope I have not failed their love and expectations. I would also
like to take this opportunity to thank our family friends: Ms. Marion Welburn, Mr. David
Welburn, and Dr. Cassie Welburn. I appreciate their time spent on proofreading my work.
Their contribution helped so much in terms of the presentation of this thesis.
I owe thanks to all these people, who witnessed me struggle along and cheered me on.
Without the help, support, encouragement, concern, trust, and love of these people, I would
not have been able to complete this thesis. This fulfilment never belongs to me alone. Instead,
it belongs to all these beloved people. By presentation of this thesis, I would like to show
them how this PhD journey has become a miracle in my life!
1
Chapter 1: Introduction
Language learners’ commitment and identity have gained scholarly attention in second
language learning research. Many of the available studies on learners’ commitment and their
identity are concerned with English, a language that is either the medium of instruction or a
mandatory subject of study in many international contexts within and beyond
English-speaking nations. In contrast, questions about learners’ commitment to, and their
identity in, Chinese language learning are newly important for policymakers, educators, and
researchers.
The absolute number of Chinese language learners remains relatively low in many
non-Chinese-speaking countries, especially in the West. However, Chinese language
programs have been growing steadily worldwide since the 1990s (Xing, 2006). Consequently,
Chinese has secured its presence among the priority taught languages in African, European,
Australasian, and wider American settings (Lo Bianco, 2011). The success of these initiatives
turns ultimately on the willing commitment of effort, time, and other resources from learners
over an extended period in which learners construct certain understandings of themselves,
others, and the world. There is widespread agreement that the nature of this commitment and
understanding varies with social contexts and learners’ characteristics. Differences between
learners of Chinese origin and those of other linguistic and cultural groups are considered by
some to be consequential for successful Chinese language programs (He, 2006; Tsung &
Cruickshank, 2011; Xing, 2006). Yet, there is little agreement about the nature of these
differences and their consequences, and there is little research to guide professional and
policy work. Moreover, there are limited studies about Chinese populations learning Chinese
2
in countries where Chinese is neither the medium of instruction nor a mandatory subject of
study (He, 2005; D. Li & Duff, 2008), such as Australia.
The research reported in this thesis investigates Chinese Australians’ commitment to
Chinese language learning and the negotiation of their ethnic identity through their Chinese
language learning. The research also explores the meaning of Chinese language for Chinese
Australians. A point of clarification is in order here. ‘Chinese’ is an umbrella term embracing
many dialects used mainly in China but also in Chinese diasporas as far as South East Asia,
North America, Australasia, and the UK. In this respect, Chinese-speaking communities use
varieties of Chinese language that is distinguished by its high level of internal diversity.
Essentially, these dialects can be categorised into seven major mutually unintelligible spoken
but orthographically similar varieties of Chinese language. W. Li (1994) summarises these
varieties and their distribution in China as:
1. Beifang (northern, the native language of more than 70% of the Chinese population):
The majority dialect, Mandarin Chinese, is a sub-variety of northern Chinese. It is
based on the speech of the capital city Beijing and has enjoyed political and cultural
significance for centuries (He, 2006). It is currently the official language of China.
2. Yue (Cantonese, 5.1% of the Chinese population): The majority of Cantonese
speakers are in Guangdong province and Hong Kong.
3. Min (Hokkien, 4.1% of the Chinese population): Hokkien is spoken in Fujian
province, Taiwan island, and Hainan island.
4. Kejia (Hakka, 3.7% of the Chinese population): Hakka speakers come from small
agricultural areas and are now scattered throughout southeast China.
5. Wu (8.5% of the Chinese population): It is spoken in the lower Yangtze River
region, including urban, metropolitan centres such as Shanghai.
3
6. Xiang (4.8% of the Chinese population): It is mainly spoken in the south central
region.
7. Gan (2.4% of the Chinese population): It is spoken chiefly in the southeast inland
provinces.
For the purpose of this study, Chinese language includes all the above mentioned dialects.
‘Chinese Australians’ in this thesis refers to Chinese immigrants and their descendants who
have gained citizenship of Australia or permanent residency in Australia. In a ‘Chinese
learning Chinese’ context such as Australia, Chinese can be described as a ‘Heritage
Language (HL)’, a language other than English that is associated with an individual’s ethnic
or cultural background (Chinen & Tucker, 2005). When Chinese Australians learn Chinese as
a HL, they are considered ‘Chinese Heritage Language Learners (CHLLs)’. These learners
have Chinese ancestry, are raised in a home where Chinese may or may not be spoken, are
primarily educated in English, and may have a variable degree of Chinese language
proficiency (He, 2006). A deeper understanding of Heritage Language (HL) and Heritage
Language Learner (HLL) will be developed in Section 1.1.
According to the Australian Bureau of Statistics (ABS) 2011 Census, Chinese ancestry,
either alone or with another ancestry, was claimed by 866,200 permanent residents, while
there was a population of 574,200 who spoke Chinese at home. The statistics lead to some
questions. For those who speak Chinese at home, why might they be committed to learning or
speaking Chinese language? What does Chinese language mean in their lives? What role does
a sense of themselves play in the process of their learning and speaking Chinese? For the rest
who do not speak Chinese at home, what has happened? Have they shifted entirely to English
or are they struggling to maintain their CHL? How does CHL loss or CHL maintenance shape
their ethnic identity?
4
In Australia, CHLLs may be mistaken for native Chinese speakers in some situations,
largely because of looking Chinese; but they may also be criticised for their incomplete
mastery of Chinese in other situations (Ang, 2001). This results in a challenge, more or less,
for Chinese Australians to find a comfortable way of living within the overlapping boundaries
of Australian society and their heritage community. The core of the challenge is the internal
balance between nationally being Australian and ethnically being Chinese (Ang, 2001)
through the tensions between nationalism and multiculturalism in contemporary Australia.
This complex Australian context will be expounded in Section 1.2.
If Chinese Australians choose either to live a lifestyle of their Chinese heritage or move
completely into the Australian lifestyle, their choices appear mutually exclusive. Alternatively,
they may choose to lead lives that include elements of both languages and cultures. As a
reflection of their challenges, some fundamental but complex questions that Chinese
Australians keep asking themselves may be: Where am I from? Where do I belong? What
does Chinese heritage mean in my life? Why do I learn Chinese? How do my decisions to use
and learn Chinese and/or English affect my identity? These questions implicate ambiguity
and complexity in Chinese Australians’ identity and their commitment to CHL learning. This
ambiguity and complexity often reflects how Chinese Australians position themselves and
how they have been positioned, sometimes unfavourably and sometimes in contradictory
ways, by their family members, their peers, their communities, and their lived worlds.
The aim of this study is to examine the interrelationship amongst Chinese Australians’
‘Chineseness’, their commitment to CHL learning, and their CHL proficiency. To clarify, this
study understands ‘Chineseness’ as a set of propensities associated with people of Chinese
ancestry, embedded within their physical attributes, rooted in their Chinese cultural history
and heritage, and emergent from their family upbringing and social learning. This study is
particularly interested in young Chinese Australian adults. This reflects several considerations.
5
Firstly, the immigrant populations of young adults with Chinese ethnic backgrounds have not
received due attention in the literature (Levesque, 2007). Secondly, Australia is a complicated
lived social world for these young adults with Chinese ancestry. Australia’s peculiar historical
and cultural background will be discussed further in Section 1.2. For this reason, the current
study identifies young Chinese Australian adults as the subjects under investigation.
Accordingly, this study seeks to answer the following research questions:
Research Question One (RQ1): Is CHL proficiency of young Chinese Australian
adults influenced by their access to various resources, the strength of their
‘Chineseness’, or both?
Research Question Two (RQ2): What does CHL mean to young Chinese
Australian adults?
The aim of Chapter One is to introduce the thesis. Section 1.1 presents definitions of
HL and HLL. Heritage Language Learners (HLLs) are distinguished from Non-Heritage
Language Learners (NHLLs). Section 1.2 depicts the particular background of the study and
locates Chinese Australians in the context of learning CHL and being Chinese ethnically in
contemporary Australia. Section 1.3 poses the researcher’s subjectivity. Section 1.4 addresses
the significance of the study. At the end of this chapter, Section 1.5 outlines the structure of
this thesis.
1.1 Definitions
‘Heritage Language’ and ‘Heritage Language Learner’ are key terms used in the current study.
In this first section of the chapter, these terms are addressed in turn. These concepts are
6
defined from various perspectives in the literature. The definitions, with different foci,
emerge from, and function in, different contexts. The definitions that best fit the context of
this study are then developed.
1.1.1 Definitions of Heritage Language
The notion of ‘Heritage Language’ has existed for a long time along with various alternative
terms such as ‘home language’, ‘mother tongue’, or ‘community language’ (He, 2008). In the
United States, HL broadly refers to immigrant languages, indigenous languages, and colonial
languages (Fishman, 2001). An inclusive definition of HL used in the U.S. context views it as
a language of personal relevance other than English (Fishman, 1989). Fishman’s definition of
HL, although elegantly simple, overgeneralises to the extent that any language could be
claimed as a language of personal relevance by someone but this language may not
necessarily function as a HL. In English-speaking countries, HL denotes a language other
than English that is associated with one’s cultural background and may or may not be spoken
at home or formally learnt in schools (Chinen & Tucker, 2005; Cho, Cho, & Tse, 1997).
Given this definition, Chinese language, currently one of the major immigrant languages in
Australia, could be considered HL for Chinese Australians.
1.1.2 Heritage Language Learners
There are many definitions in the literature for the term ‘Heritage Language Learners’. These
definitions vary in terms of contexts and foci such as place, HL proficiency, heritage
membership, and identity. All definitions are valid and of value for particular contexts.
However, as HLLs cover a very heterogeneous population, there is no universal definition
capable of embracing all individuals under the heading of HLLs in all situations. The optimal
7
definition will function in a specific context where this definition emerges. Accordingly, the
following discussion develops a definition useful for this study.
HLLs in the United States are defined as individuals who are members of a community
with linguistic roots in a language other than English and who are learning the language of
that community (Cho, 2000). The main determinants of being a HLL in this definition are
association with a HL community and identification of membership in that community. By
this definition, proficiency in the HL is not taken into account. This definition may apply to
HLLs in a community that has a strong sense of HL roots. Where there are limited numbers
of HL speakers who are struggling to establish a community and are striving to reverse
language shift, this definition might be overly narrow.
Other definitions of HLLs extend to the ancestral language of an individual or group,
regardless of whether that language is still used at home (Fishman, 2001; Noels, 2005). HLLs
are individuals who have the desire to learn an ancestral language that was the language
spoken by previous generations of their families and that is not the language of the dominant
society (Cummins, 1998; Noels, 2005). Like the previous definition, this definition does not
take account of HL proficiency. However, it does highlight desire and intentional
commitment to HL learning.
Some definitions have emerged to take account of both heritage membership and HL
proficiency. One definition includes both learners who have either some degree of bilingual
proficiency or a cultural connection to the HL, and learners with a heritage motivation who
perceive a cultural connection that is more distant than that of first- or second-generation
immigrants (Van Deusen-Scholl, 2003). In an English-speaking country, HLLs are defined as
learners who a) are raised in and belong to a home and community where a language other
than English is spoken; b) speak or at least understand the HL; and c) are to some degree
bilingual in English and the HL (Valdés, 2001). These definitions indicate that both HL
8
proficiency and membership in a heritage community serve as criteria for determining HLL
status. This is a point of distinction from the definitions of Cho (2000), Cummins (1998),
Fishman (2001), and Noels (2005). However, the assumptions that HL is necessarily used at
home or in a heritage community and that HLLs have certain level of HL proficiency are
problematic.
Some definitions highlight the role of identity in the process of HL learning. HLLs are
those who have identity and/or linguistic needs with regard to HL learning that are related to
their heritage background (Carreira, 2004). In other words, their commitment to learning the
HL arises from a desire to connect more fully with their heritage or understand their identity
more deeply.
Considering the previous definitions, the term ‘Heritage Language Learner’ is used in
the current study to describe Chinese Australians who a) are Chinese learners with a cultural
and heritage connection to a Chinese language and its community; b) are educated primarily
in English; c) may or may not speak or understand a Chinese language; and d) may be
bilingual in a Chinese language and English language. This is an inclusive definition. It is
useful for the present study because it can describe the complexity of the Chinese Australian
population. Chinese Australians comprise a heterogeneous group ranging from those fluent in
Chinese language to those with no Chinese language at all. On one end of the spectrum, they
may be those who grew up with a Chinese home language other than English. On the other
end of the spectrum, they may be either third generation or further removed or inter-racial
adoptees with no exposure to CHL but feelings of cultural connection to CHL and Chinese
identity. The purpose of learning for this diverse population varies from CHL maintenance to
CHL development respectively. Nevertheless, CHLLs are very different from Non-Heritage
Language Learners.
9
1.1.3 Heritage Language Learners and Non-Heritage Language Learners
Heritage Language Learners are a particular class of language learners. Although they may
share certain characteristics and experiences with non-Heritage Language Learners, they
must nevertheless be conceptualised as language learners who possess distinct features from
NHLLs (Lo-Philip, 2010), in terms of their background knowledge of HL, familiarity with
the heritage culture, and forms of desire for HL learning (Campbell & Rosenthal, 2000).
Chinese and non-Chinese Australian learners of Chinese, for example, are likely to be
significantly different in at least some of these aspects. These aspects are discussed below.
Firstly, HLLs and NHLLs may differ in language proficiency. Due to limited exposure
to the HL and culture, HLLs may differ from First Language Learners of the same language
in that they often do not fulfil the linguistic competency in the HL at the same level of First
Language Learners (Carreira, 2004). However, there is a general agreement on the growing
evidence that even limited childhood language experience can help in the acquisition of that
language later in life (Au, Knightly, Jun, & Oh, 2002). In line with this perspective, HLLs are
more likely than Second Language Learners or Foreign Language Learners to have had some
prior exposure to the language, though often limited to early childhood, and therefore they
may become proficient in the language in ways that Second or Foreign Language Learners
are not and may never be. Prior linguistic and cultural knowledge of HL makes HLLs more
likely to achieve better language skills than Second or Foreign Language Learners,
particularly in speaking and listening (Ke, 1998; Shen, 2003; Xiao, 2006). On the other hand,
given differences in acquisition contexts, HLLs may have wide gaps in HL knowledge and
skills that NHLLs do not have (Carreira, 2004). For example, HLLs may need to connect oral
language skills developed through prior exposure to the language with the written form of the
language. In contrast, NHLLs often need to develop oral and written skills simultaneously at
the outset. Therefore, unlike NHLLs, HLLs may be quite comfortable with speaking and
10
listening but may often possess limited skills in reading and writing (A. Louie, 2004; Xiao,
2006). These gaps may prevent them from performing certain kinds of reading and writing
communicative tasks at a level commensurate with their listening and speaking skills (J. B.
Webb & Miller, 2000).
Secondly, most distinctions between HLLs and NHLLs are concerned with the way
that many HLLs deal with issues of identity (Peyton, Ranard, & McGinnis, 2001). HLLs
appear to have a multi-faceted identity as someone who is both similar to and different from
members of the heritage culture since they are socio-historically connected with the heritage
culture and yet experientially displaced from it (He, 2004). To a greater extent than NHLLs,
HLLs are likely to be driven by identification with the intrinsic cultural, affective, and
aesthetic values of the language (He, 2006), and by the goal of gaining greater access and
depth to their ethnic identity (Chinen & Tucker, 2005; He, 2006). Therefore, HLLs often
bring with themselves a set of ambiguities and complications in identity, which may be
absent in NHLLs (He, 2006).
Thirdly, HLLs and NHLLs also differ in terms of commitment to language learning.
HLLs often have different desires for studying the language than NHLLs for they are seeking
greater understanding of their heritage culture or stronger connection with their family and
community (Draper & Hicks, 2000). HLLs are committed to HL to communicate with family
members, to participate in the life of their HL community, to feel connected to their roots, to
contribute to the preservation of their beliefs and practices, or to overcome feelings of being
an outsider in their heritage community (Carreira, 2004). Of course, HLLs may also be
committed to HL learning for additional reasons such as to fulfil a language requirement as
part of a course or to pursue a career. Therefore, the commitment of HLLs to HL learning is
likely to be derived not only from pragmatic or utilitarian orientations, as tends to be the case
with NHLLs, but also from intrinsic cultural connections to the HL.
11
To help understand the nature and dynamics of Chinese Australians qua CHLLs, the
next section will introduce the complex cultural, social, and historical Australian context,
within which Chinese Australians negotiate tensions between nationally being Australian and
ethnically being Chinese.
1.2 Background
The compression of time and space by the innovation of information technology and the
intensification of transnational migrations of people are creating a more or less borderless
world. Mobility and connections are no longer as restricted to local regions as before. The
world is thus moving from territorially distant and mutually distinct spaces towards
multi-dimensional space of overlapping layers. There is a transition underway from a socially
distinct, culturally homogeneous, and politically sovereign world of nations to an
interconnected and intermingled one (Ang, 2001). People are intersecting and interacting in
such a world where they are on the move as they have never been before. This migration has
shaped the present time as an era of growing diversity. It has introduced new cultures and
languages to destination countries, and rapidly increased the cultural and linguistic diversity
of modern societies.
Against the background of regional and global migration as well as transnational
communications and interactions, the current section explores and understands the cultural
and linguistic diversity in Australia, a space of complicated entanglement. This section also
introduces the diverse cultural and linguistic backgrounds, needs, aspirations, and potentials
of Chinese Australians, who are faced with the challenges and opportunities in Australia, a
place of “togetherness-in-difference” (Ang, 2001, p. 17).
12
1.2.1 Cultural and language policies in Australia
Australia was already inhabited and owned by Aboriginal and Torres Strait Islander peoples
before the history of Australian immigration started. The early interactions between these
peoples and those from nearby islands were documented in archaeological work (Macknight,
1976): Fishermen from Makassar in the southern Celebes (Sulawesi) frequently visited the
coast between the Cobourg Peninsula and the Sir Edward Pellew Group in the early
eighteenth century. Given these early interactions, it is not surprising that some Aboriginal
languages contain a substantial number of Makassan words (Macknight, 1976).
The first cohort of British immigrants landed on this continent at Botany Bay on 26
January 1788. Since this historical moment, people with different cultures and languages
have immigrated to the continent from different parts of the world. The immigrant population
of Australia continues to grow rapidly. As shown in the 2011 Australian Census, three out of
ten Australian residents were born overseas, compared with one out of ten and two out of ten
in 1947 and 1971 respectively (Castles & Davidson, 2000). A wealth of linguistic and cultural
resources is available in Australia through its culturally and linguistically diverse population.
This diversity will continue to be a characteristic of future Australia (Holmes, Hughes, &
Julian, 2007).
The wealth of cultural diversity in Australia has been valued and supported by the
nation through its implementation of multiculturalism as a political strategy to accommodate
and include ethnic minorities (Ang & Stratton, 2001). However, in Australia, tensions exist
between a ‘unified nation’ and a ‘multicultural one’ (Holmes et al., 2007). A narrative gloss of
these tensions reflected by Australia’s immigration policy over complex processes could
include (Holmes et al., 2007): 1901 to the early 1970s (White Australia policy); 1950s to the
early 1970s (dismantled White Australia policy with assimilation); and the early 1970s
onwards (multiculturalism policy).
13
One of the first pieces of legislation enacted by the new national parliament upon
Federation was the Immigration Restriction Act 1901. This act was meant to manage
competition and labour disputes in the goldfields and growing tensions around Australian
nationalism. The Act signified the commencement of the so-called ‘White Australia Policy’
and formed its legislative basis. The White Australia Policy, including the 1901 Act and the
subsequent acts to strengthen it, intentionally restricted non-white immigration to Australia
from 1901 to the early 1970s.
During World War II, the arrival of thousands of refugees in Australia with the
Japanese advance; the service of numerous Australian Aborigines, Torres Strait Islanders,
Papua New Guineans, and Timorese in the frontline to defend Australia; and Australia’s
vulnerability in the Pacific War because of its small population brought Australia’s racially
discriminatory immigration and political rights into focus. After the conclusion of World War
II, the White Australia Policy was dismantled in stages by successive governments with the
encouragement of the first non-British and later non-white immigrants. Wartime service also
gave many Indigenous Australians confidence in the claim of their equal rights upon return to
civilian life.
In the post-war period, assimilation became the dominant policy in Australia, where
assimilation refers to the process by which indigenous people and immigrants were expected
to relinquish their culture and become indistinguishable from the Anglo-Australian dominant
group (Holmes et al., 2007). The ideal or illusion of national-cultural homogeneity was that
all indigenous people and immigrants would be incorporated smoothly within the dominant
national culture, leaving their original cultures happily behind. The dominant culture was
considered the national culture on which national identity was to be constructed. Ethnic
cultures were defined by the dominant culture as ‘different’. The national identity was
homogeneous and it denied ethnic cultures that threatened it.
14
The ‘assimilation’ model was a political strategy to balance inclusion and exclusion.
On the one hand, the model incorporated ethnic cultures only insofar as they could contribute
to the ideological discourse of cultural diversity in Australia. On the other, this model, per se,
marked immigrants as ‘others’ to the dominant group. As such, assimilation can never be
fully successful. The acquired, rather than inherited or ascribed character of cultural traits
gained in the process of assimilation turns the assimilating subjects into less than ‘real’ and
still inferior Australians (Ang, 2001). This led to the introduction of multiculturalism, which
marked the dismantling of the White Australia Policy in the early 1970s.
In the 1970s, the Australian Ethnic Council adopted a formal statement against
assimilation and immigration policy moved from assimilation to multiculturalism (Holmes et
al., 2007). Since then, multiculturalism has become a key element of government cultural
policy and the description of Australia as a multicultural nation has become commonplace
(Ang & Stratton, 2001; Holmes et al., 2007). However, a distinction between multicultural
Australia and multiculturalism in Australia has been made in terms of the distinction between
practice and policy (Hodge & O’Carroll, 2006). ‘Multicultural Australia’ refers to the
cultural diversity in which interaction between different cultures is in process (Holmes et al.,
2007). It is a shifting and dynamic interweaving of cultures and diversities actually happening
and existing in Australia. In contrast, ‘multiculturalism’, as a centrepiece of national cultural
policy, refers to the management of cultural diversity (Ang & Stratton, 2001) but within
certain and well-demarcated limits without disturbing or threatening national unity (Ang,
2001). This policy expects that all members of society have equal rights, regardless of ethnic
background (Holmes et al., 2007). It is an ideology to prescribe what should be happening or
speculate what could be happening by the acceptance and promotion of multiple cultures in
the demographic context of a specific place, such as organisations, communities, cities or
nations.
15
Multiculturalism, as a national policy in Australia, has a complicated history. Firstly,
the adoption of multiculturalism functions to discard the racist past in Australia. The
exclusionary and homogeneous White Australia Policy was replaced by multiculturalism, a
discourse in favour of pluralism and heterogeneity and recognition of valuable cultural
diversity. Secondly, multiculturalism functions to convince Australians of the public fiction
that they live in a harmonious and inclusive society. Multiculturalism becomes an ideological
discourse addressed to Australians to promote unity within diversity. Thirdly, the ideological
discourse of multiculturalism has a political orientation. It functions to serve the policy of the
government. Politicians have announced with pride that Australia is one of the most
successful multicultural societies in the world, though support for the creation of
multicultural Australia has always been less than whole-hearted (Ang & Stratton, 2001).
Language policies in Australia largely reflect Australia’s complicated cultural history.
During the ‘White Australia’ era, English was assumed to be the national and only necessary
language. Immigrants were expected to learn English, leaving their native languages happily
behind. There was no concept of multiculturalism as there is today. In the 1950s and 1960s,
the White Australia Policy was gradually dismantled. Although the period saw little change
from the traditional and classical ‘English only’ approach, there were remarkable
improvements in teaching English as a Second Language (ESL), in order to enable
immigrants arriving in Australia from different parts of the world to assimilate more readily.
The Adult Migrant Education Program was very innovative in its approach to English
teaching and learning to support ‘assimilation’. However, White culture was still the
dominant core surrounded by multiple minority ethnic cultures.
From the dysfunction of the White Australia Policy in the early 1970s (Holmes et al.,
2007) to the mid 1990s, Australia probably led the English-speaking world in systematic
language policy making (Ingram, 2000). The 1970 review, Teaching Asian Languages and
16
Cultures in Australia, made the first attempt to direct language policy towards Asia
(Commonwealth Advisory Committee, 1970). The review reflected Australia’s growing
political realignment towards Asia, but paid little attention to the economic importance of
Asian languages to Australia. Following the review, there was a sudden realisation that
Australia, almost accidentally, imported numerous languages that would be wasted if no
language policy enabled them to survive and to be used; and that many children entered
schools speaking their HL and needed an opportunity to commence their education in their
HL. Accordingly, there were many unsystematic attempts to reform language education. The
Department of Education advocated that there were “strong educational and social reasons for
migrant children” entering schools with “inadequate” English proficiency to continue the
learning of their HL (1976, p. 35). Guided by this initiative, funds were allocated to schools
to establish HL programs and to community schools to support the teaching of HL.
Another important document during the multicultural period of the 1970s was the
Galbally Report released by the Australian Federal Government in 1978. It took the position
that “every person should be able to maintain his or her culture without prejudice or
disadvantage and should be encouraged to understand and embrace other cultures” (Galbally,
1978, p. 4) and further that “cultural and racial differences must be reflected in educational
programs designed to foster intercultural and interracial understanding” (pp. 104-105). These
fundamental initiatives, though unsystematic, were to encourage enrichment and diversity
within an adherence to certain core values in the society. These initiatives also provided the
foundation for the development of the first National Policy on Languages (Lo Bianco, 1987)
in the English-speaking world.
On its release, the National Policy on Languages became the standard against which
State and Territory policies could be compared. This policy renewed language policy in
Australia, stressing such aspects as the support for the maintenance of HLs, Aboriginal and
17
Torres Strait Islander peoples’ languages, and the extensions to programs for teaching ESL. It
made a wide range of recommendations, taking account of the breadth of the social context in
which the policy was placed (Ingram, 2000). Moreover, the policy focused on the labour
market and the ways in which tackling adult illiteracy levels, extending English proficiency,
and teaching ‘trade languages’ would benefit Australia’s economic performance (Lo Bianco,
1990). To a large extent, the National Policy on Languages culminated the phases of
pluralistically oriented language policy of the 1970s and 1980s (Lo Bianco, 2000). However,
the policy lacked rigorous frameworks to address teacher education and teacher supply.
Funding distributed to stimulate the teaching of other languages was mainly through projects.
Unfortunately, when such funding dried up, the long-term effect of the projects was
questionable. There was scant on-going monitoring and evaluation of the policy and its
programs to ensure that the policy was continually evolving in response to emerging needs.
These deficiencies of the National Policy on Languages contributed to the development of
the Australian Language and Literacy Policy released by the Department of Employment,
Education, and Training (DEET) in 1991 (Ingram, 2000).
The Australian Language and Literacy Policy explicitly claimed to be derived from
and closely influenced by the National Policy on Languages (DEET, 1991). The 1991 policy
had the basic position that “Australian English” was the “national language”, but the national
priority languages, such as Asian languages, helped to “enrich the intellectual and cultural
vitality” of Australia and secure the “future economic well-being” of Australia (DEET, 1991,
pp. iii-iv). Accordingly, it foregrounded English and Asian languages and tied the skills of
these languages to education, employment, and trade. Despite the value attached to
multiculturalism and the maintenance and development of Aboriginal and Torres Strait
Islander peoples’ languages and HLs, the policy reflected the prevailing economic rationalism
that had dominated Australian language policy thinking and making since the late 1980s
18
(Ingram, 2000). In other words, it placed more emphasis on economic reasons for language
education than cultural or multicultural importance. As such, it focused more on language
skills and problems than language rights and resources.
As a complement to the Australian Language and Literacy Policy, another policy,
Asian Languages and Australia’s Economic Future, released by the Council of Australian
Governments (COAG) in 1994, continued to highlight the importance of teaching and
learning of Asian languages, in particular the four priority languages of Japanese, Chinese,
Indonesian, and Korean. Similar to the previous policy, the 1994 report emphasised the
economic value of Asian language skills as tools to facilitate competitive international trade.
The report set proficiency targets to be achieved in schools, encouraged early commencement
of learning Asian languages in Year 3, and recommended that a nationally agreed minimum
skill level be specified for Asian language teachers (COAG, 1994). However, the report was
criticised for being impractical in its targets and assumptions (Lo Bianco, 2000). An
under-supply of qualified Asian language teachers stood against its implementation. A further
criticism concerned the over-emphasis on economic reasons for language learning and the
neglect of cultural and intellectual values (Ingram, 2000).
Building on previous experiences, the Australian Curriculum, Assessment, and
Reporting Authority (ACARA) released the Shape of Australian Curriculum: Languages in
2011 to guide the development of language curriculum in Australia. The language curriculum
recognised the importance of a capability in languages in economics, diplomacy, trade,
cultural exchange, and national security. The language-specific curricula will be developed
for different languages and different groups of language learners that include First, Second,
and Heritage Language Learners. Within each of these groups there are differences in
proficiency in using the target language, with the span of language experiences of HLLs
particularly wide and the affiliations with their HLs particularly diverse (ACARA, 2011).
19
Being aware of these differences, the language curriculum designed the aims, pathways,
programs, hours of study, and achievement standards for language learning. Therefore, the
language curriculum is expected to improve language education in Australia, increase
investment in research and in public services dealing with languages, support ESL programs
for immigrants and the intergenerational maintenance of immigrants’ HLs, ensure recognition
of Aboriginal and Torres Strait Islander peoples’ languages, and facilitate language revival
initiatives for indigenous languages at risk of extinction (ACARA, 2011).
Being conscious of the “unstoppable” transformation of the Asian region into the
economic powerhouse of the world and the “gathering pace” of this transformation, the
Australian Federal Government released the Australia in the Asian Century White Paper in
October 2012, aiming to help “seize the economic opportunities” and “manage the strategic
challenges” of this development (Australian Government, 2012, p. ii). To broaden and deepen
the people-to-people links between Australia and the Asian region, particularly to strengthen
the nation’s principal relationships with China, India, Indonesia, and Japan, the White Paper
urged Australians to become more “Asia literate” and “Asia capable” (Australian
Government, 2012, p. iii). Detailed strategies for studies of Asia will become a core part of
school education and all schools will engage with at least one school in Asia to support the
teaching of a priority Asian language. These languages are Chinese (Mandarin), Hindi,
Indonesian, and Japanese. All Australian students will be encouraged to study and will have
access to at least one priority Asian language. Measures to track how Australian students are
increasing their knowledge of Asia will be developed. Industries and communities will be
encouraged to increase their understanding of the benefits of learning priority Asian
languages so that the demand for Asian language studies will be boosted. Australian
universities will be supported to increase the number of students who undertake Asian studies
and Asian languages as part of their university education, and will be encouraged to establish
20
an exchange arrangement involving transferable credits with at least one major Asian
university. Funding will be available, through the Australian Research Council and other
mechanisms, to strengthen research and teaching links between Australian and Asian
institutions. In brief, support and encouragement of Asian languages through school
education, university teaching and research, and industrial and community engagement are
key elements of the White Paper. The White Paper not only addresses the economic demand
for Asian languages but also makes the cultural diversity available in Australia through its
support and encouragement of language pluralism (Australian Government, 2012).
To sum up, the current model of multiculturalism and language policy in Australia
conceptualises ethnicity, national identity, and globalisation. This is a model of ethnicity as
flux and flow, national identity as a heterogeneous construction, and globalisation as a set of
ongoing processes that challenge the boundaries and sovereignty of the nation-state (Holmes
et al., 2007). This model reconfigures Australian nationalism, shifting from a racially
exclusionary form to an inclusive and open-ended form. This model offers a view from the
margins of Australian society (Holmes et al., 2007). It challenges the dominant group’s views
on ethnicity and nationalism. It is a vision of negotiated social cohesion predicated on the
acceptance of cultural difference as an integral part of contemporary Australia. It is within
these complex cultural and language contexts that Chinese Australians negotiate their
identities and learn CHL, willingly or unwillingly.
1.2.2 Chinese immigrants in Australia
As immigrants from every corner of the world continue to seek entry into Australia,
Australian cultural and language policies have to deal with the question of how to manage
proliferating differences and diversities. These policies have a remarkable impact on
21
immigrants in Australia. As one of the earliest non-White settlers, Chinese immigrants and
their descendents have formulated various forms of life politics in response to the shifting
cultural and language policies in Australia. The following discussion focuses particularly on
Chinese immigration in Australia, looking at both the impact of cultural and language policies
of Australia and that of the historic situations of China.
The Chinese people have a long history of emigration overseas. One of the earliest
emigrations can be dated back to the Ming Dynasty when He Zheng (1371-1435) became the
envoy of the Ming Dynasty. He sent people to explore and trade in the South China Sea and
in the Indian Ocean. The great Chinese diaspora began in the 19th century, the boom age of
European colonialism. Due to the lack of labourers in many colonies and the poverty brought
by population pressure, natural disasters, heavy taxation, destructive rebellions, and
widespread opium smoking in southeast China, many Chinese emigrated from southeast
China to work in Southeast Asia, where they had established earlier links since the Ming
Dynasty. Many other Chinese migrated to countries in North America and Australasia, where
there was a great demand for labour in gold mining and railway construction. On the one
hand, widespread famine and poverty in southeast China impelled many Chinese to work in
these countries so that they could earn more money to improve the living conditions of their
family members and relatives. On the other hand, the demand for labour in these countries
was addressed. Since the early 19th
century, Chinese emigration has been directed primarily
to North America and Australasia.
From the very beginning of Australian immigration history, links with China were
established when several ships dropped off their convict load in Australia then sailed for
southeast China to pick up goods for their return to England (Cushman, 1984). The earliest
documented Chinese immigration to Australia dates back almost two hundred years, with
Mak Sai Ying being the first recorded Chinese settler in 1818 and with the first large group of
22
immigrants arriving from southeast China in October 1848 (Cushman, 1984). From this time
onwards till the late nineteenth century, Chinese immigration was seen as part of a solution to
labour shortages in Australia (Cushman, 1984).
In the mid 1850s, many more Chinese gold-seekers arrived at the various diggings.
Large numbers of early Chinese immigrants worked on goldfields. Others started to open
stores and became merchants and hawkers. Over the 1860s and 1870s, fishing and fish curing
industry were commonly operated by Chinese immigrants in Sydney, supplying dried fish to
Chinese people throughout New South Wales as well as Victoria. Stores and dormitories run
by Chinese people soon developed to support the miners on the fields as well as those on
their way to the diggings or back to China. The attraction of gold and competition in gold
mining aroused the resentment of European diggers towards Chinese diggers due to cheaper
Chinese labour. At the same time, attempts to maintain Chinese as indentured labourers were
difficult because some Chinese diggers often deserted their employers for more lucrative
individual gold mining. These problems regarding the Chinese population drew the
government’s attention and resulted in restrictive anti-Chinese legislation in the late 1850s
and the early 1860s. The objection at this time to Chinese immigration was economic
competition and cultural differences between the white and Chinese ways of life, rather than
feelings of racial superiority (Choi, 1975).
These restrictive measures effectively reduced the inflow of Chinese. But with a labour
shortage in the gold-mining industry and the dramatic decrease of Chinese immigrants, the
government repealed the restrictive acts in the 1860s (Choi, 1975). The Chinese population
continued to decrease for a short while after the repeal but then rapidly increased. Again, the
government became concerned with the growing number of Chinese in the mining areas and
feared that conflict between Chinese and European diggers and breaches of law and order
might occur. This resulted in restrictive Acts in the 1880s.
23
Despite the shifting immigration laws, Chinese immigrants managed to remain and
often prosper (Choi, 1975). While the major body of Chinese immigrants was digging on the
goldfields, many others endeavoured to try other ways of earning a living very soon after
their arrival. By the 1890s, Chinese people were represented in a wide variety of occupations
as scrub cutters, cooks, tobacco farmers, market gardeners, cabinet-makers, storekeepers,
laundry workers, and drapers, though by this time the fishing industry seemed to have
disappeared. They ran stores, import trade, and several Chinese language newspapers. They
were also part of an international community involved in political events in China such as
sending delegates to the Peking Parliament and making donations at times of natural disaster.
By the 1890s, the colonial parliaments had placed a series of restrictions on the
migration of ‘coloureds’ in general and Chinese in particular (Jones, 2005). Continuing
competition and labour disputes in the goldfields, as well as Australian nationalism, created
an environment of racial antagonism during the second half of the 19th
century. This led to the
new Federation’s Immigration Restriction Act 1901, the so-called White Australia Policy. It
was this policy that severely curtailed the development of the Chinese communities from the
late 19th
century onwards and led to the steady decline of Chinese population in Australia
(Jones, 2005). Continued discrimination, both legal and social, reduced the occupational
range of Chinese people. Gardening became one of the major occupations. Chinese
immigrants, throughout the first 30 years or so of the 20th
century, relied on successful
Chinese merchants to assist them to negotiate with the Immigration Restriction Act
bureaucracy. It was the rise of a new generation of Australian-born Chinese people, combined
with new immigrants that the merchants and others sponsored, that prevented the Chinese
population from dramatically declining (Choi, 1975).
At the beginning of World War II, Australian-born Chinese began to outnumber
China-born Chinese for the first time. However, with the Japanese invasion in China and
24
other Asia-Pacific regions, large numbers of refugees boosted the numbers of arrivals from
China again (Jones, 2004). Some were Chinese crewmembers who refused to return to
Japanese-held areas. Others were residents of the many Islands of the South China Sea
evacuated in the face of the Japanese advance. Still others were of Australian birth and were
able to leave Hong Kong for Australia on the approach of the Japanese (Jones, 2004).
After the conclusion of World War II, the White Australia Policy was gradually
dismantled. At the same time, restaurants began to replace gardening as the major source of
employment and avenue for bringing in new Chinese immigrants. These changes brought
about the end of the dominance of south China in the link between China and Australia that
had existed for over 100 years. For the first time numbers of Chinese immigrants from
non-Cantonese speaking parts of China significantly increased. During the decades from the
1950s to the 1970s, Australia’s restrictive migration policy was gradually relaxed before its
formal abolition. However, following the establishment of the People’s Republic of China in
1949, movement from China was most difficult, with a small number of Chinese arriving
each year. There was almost an absence of Chinese emigration from the 1950s to the 1970s.
During this period, China placed severe restrictions on the movement of its citizens.
From 1973, the White Australia Policy was for all practical purposes defunct. In 1975,
the Australian government passed the Racial Discrimination Act, which signified the end of
the White Australia Policy. Multiculturalism started as official policy in Australia. Only a few
years later, China opened its door again to the world. The end of the White Australia Policy
and the commencement of the Chinese Opening-up Policy saw new arrivals from China in
Australia. Since then, immigrants from the Chinese Mainland arrived in increasing numbers.
From the 1970s onwards until the end of the twentieth century, there were two significant
booms of emigration from China, in the late 1970s and the early 1990s respectively. In the
late 1970s, emigration restrictions in China were eased as a result in part of the Opening-up
25
Policy. More liberalised emigration policies were enacted to facilitate the legal departure of
increasing numbers of Chinese who joined their overseas Chinese relatives and friends. At the
same time, the ‘Four Modernisations Program’ that required Chinese students and scholars,
particularly scientists, to attend foreign education and research institutions brought about
increased contact with the outside world. Anyone who had the necessary economic resources
could apply for permission to study abroad. Other political events in China also promoted
immigration. In 1984, Britain agreed to transfer sovereignty of Hong Kong to China, and in
1989, a student movement broke out at Tiananmen Square. These two historical events
triggered a lack of confidence in the Chinese government and accelerated another wave of
migration to the United Kingdom, Australia, Canada, USA, and other parts of the world in the
early 1990s. The wave calmed after Hong Kong’s transfer of sovereignty in 1997.
Australian immigration policy has been transformed from ‘White Australia’ to
‘multiculturalism’. Almost at the same time, Chinese diplomatic policy has changed from
‘closing-down’ to ‘opening-up’. These policy changes saw new arrivals from China to
Australia. New institutions were established for these arrivals and old ones, such as the
Chinese Chamber of Commerce, revived. Chinese language newspapers were once again
published. Shortwave and longwave radio channels began to broadcast in Cantonese and
Mandarin. The Australian public broadcaster SBS also provided television and radio
programs in both languages. The Chinese language, mainly Mandarin and Cantonese, became
available as a subject in some secondary schools. Community language schools started to
operate on weekends. Several Chinese Australians have received the Order of Australia award
and there are current Chinese Australian representatives in both State and Federal parliaments
(Jones, 2005).
The population with Chinese ancestry keeps increasing in Australia. Today, Chinese are
the third largest group among all immigrants in Australia, just behind people from the United
26
Kingdom and New Zealand (ABS 2011 Census). With Chinese immigration, Chinese
language was brought to Australia. It is now the most widely spoken language other than
English at home (ABS 2011 Census). The following section provides some background
knowledge about Chinese language in Australia.
1.2.3 Chinese language in Australia
Chinese immigrants have brought and continue to bring Chinese language to Australia. With
the increase of Chinese immigrants in Australia, the population that speaks Chinese at home
has also increased. Where it was 401,357 in 2001 (ABS 2001 Census), it was 574,200 in 2011
(ABS 2011 Census). This makes Chinese the most common of all languages other than
English brought to Australia. Three percent (3.0%) of the national population speaks Chinese
at home (ABS 2011 Census).
On the international stage, China continues to speak to the world about its sheer
vastness, its huge population, its fascinating history and culture, its idiosyncratic Chinese
socialism, and its rapid economic growth. There has been growing interest in China as an
emerging international power and a potential cooperative business power (Zhang &
Slaughter-Defoe, 2009). For Australia, China is now a regional neighbour and trade partner. It
has been recognised that the development of the relationship with China and mutual benefits
of such require a solid pool of Australians in a range of sectors who can understand China
deeply and speak Chinese well (Orton, 2008). This strategic viewpoint was stressed in the
Australia in the Asian Century White Paper released by the Australian Government in
October 2012.
The teaching and learning of Chinese began two decades ago to produce
Chinese-speaking graduates from Australia’s schools to serve the country’s economic,
27
cultural, and political interests (Sturak & Naughten, 2010). However, the share of Australian
students learning Chinese is relatively small and has fallen in recent times. In recognising this
decline as well as the need to build a sound knowledge of Chinese in schools, the Australia in
the Asian Century White Paper advocates that the languages component of the Australian
Curriculum will enable all students to learn a language other than English, with a curriculum
for Chinese being one of the first in development. Compared to the previous situation,
ambitious targets have been set, especially in curriculum design, assessment mechanisms,
textbook development, and in the diversity of Chinese programs offered.
There has been increasing attention to teaching and learning Chinese as a Foreign
Language in Australia, not only due to the effort of the Australian Government but also
because of the recent sponsorship of China’s National Office for Teaching Chinese as a
Foreign Language (abbreviated Chinese name: Hanban). China has recognised the
importance of promoting Chinese language and culture to the international community, and
began to establish non-profit institutions in 2004. These institutions were given the name
‘Confucius Institutes’ or ‘Confucius Classrooms’. They aim to promote Chinese language and
culture in foreign countries. Since 2005, nine Confucius Institutes and 11 Confucius
Classrooms have been established in Australia, offering diverse Chinese programs at various
levels outside the regular education systems.
Chinese is now one of the major languages taught, both in the education system and in
other Australian settings. Interestingly, the majority of Chinese learners in Australia are
Chinese Australians. At senior secondary level, an overwhelming proportion (90%) of
Chinese learners are Chinese Australian students (Orton, 2008). In higher education, the
positive gains in Chinese language learners’ enrolments largely reflect the growing number of
young Chinese Australian adults (McLaren, 2011). In these contexts, Chinese language is
taught and learnt as a HL rather than a Foreign Language. This raises the question as to how
28
well Chinese Foreign Language programs match the needs of CHLLs. Teaching and learning
Chinese as a Foreign Language is becoming a topic of interest in Australia. However,
teaching and learning Chinese as a HL has not been effectively addressed in relation to its
unique CHLL population (McGinnis, 2008). Though Chinese language is the most common
HL with a growing number of CHLLs in Australia, CHL has received little attention in terms
of theory-building thus far (Wang, 2007). Very little is known about the nature and the
dynamics of CHL learning as well as the rate and route of CHL maintenance and
development (He, 2006). The study reported in this thesis seeks to fill the gaps in the
literature and to address the associated practical issues.
1.3 Researcher’s subjectivity
Taking account of the weight of the quantitative component in this mixed methods study, ‘the
researcher’, instead of ‘I’, was used throughout the thesis in line with the convention of
reporting quantitative research. However, equally mindful of qualitative research conventions,
it is also worth addressing the researcher’s subjectivity at this point. As a Chinese citizen and
a native Mandarin speaker, the researcher came to Australia to pursue his doctoral degree.
Recognising the transformation of China into the economic powerhouse of the global
economy and the accelerating pace of this transformation, the researcher was interested in the
cultural and social impact of this transformation on Australia, a country that has identified
China as a strategic partner in the Asia-Pacific region. He is convinced of the role that
Chinese language can play in strengthening this partnership to mutual benefit. There is an
additional consideration with regard to researching the Chinese Australian population.
Because the researcher is culturally closer and physically more similar to Chinese Australians
than any white scholar would be, he may achieve better entrée to and rapport with the
29
Chinese community in Australia and access to data that may not be available to white
scholars. However, this relatively easy access and affinity with the research participants do
not necessarily lead to biased interpretation of the data or taking the data for granted. As an
international student from China, the researcher was relatively confined to the university life
and thus removed from the more complex lived worlds of Chinese Australians. Such social
difference offered the researcher some distance from the participants’ experiences as an
outsider. This combination of ‘insider’ and ‘outsider’ status will be revisited in Chapter Four.
Beyond these contextual and personal reasons, the researcher investigates how Chinese
Australians negotiate their identities and capitalise on various resources through their CHL
learning.
1.4 Significance of the research
This study has theoretical, methodological, and practical significance. Theoretically, HLLs’
ethnic identity and their commitment to HL learning have received attention from different
perspectives, including but not limited to, classical social psychological scholarship and
poststructuralist scholarship. Both schools offer meaningful perspectives in certain contexts,
but receive critiques from other contexts, as will be discussed in Chapter Two. Relevant
issues associated with CHLLs seemed to have been overlooked in the literature. These issues
are particularly salient for Australia, a complex social place with an idiosyncratic historical
and cultural background. The current study makes a theoretical breakthrough to examine
Chinese Australians’ commitment to CHL learning and their ethnic identity in CHL learning
through Bourdieu’s sociological theory to reconcile the oppositions between classical social
psychological and poststructuralist perspectives. This will be discussed in detail in Chapters
Two and Three. Methodologically, empirical studies using Bourdieu’s sociological theory to
30
address language learning-related problems have often adopted qualitative research methods.
This mixed methods study combines a quantitative approach and a qualitative approach to
add depth and scope, which resonates with Bourdieu’s own methodological thinking.
Particularly, the study makes a methodological breakthrough: A set of instruments was
developed and validated to quantify Bourdieu’s key concepts of capital and habitus within
certain social fields, as will be discussed in Chapter Four. Practically, the benefits of HL, for
both individuals and society, have been well documented in recent studies (C. Baker, 2003;
Cho, 2000; J. S. Lee, 2002; McGinnis, 2005; Shin, 2010; Tse, 2001) and current language
policies. The current study attempts to understand Chinese Australians’ personal, social, and
cultural engagement in CHL learning, which may make Chinese Australians better
understand themselves within the social world that they live in and make them better
understood by the Australian society; and further promote enhanced intercultural relations,
adaptable thinking, and mutual understanding within multicultural Australia. As such, this
study resonates with the political discourse in the current language and cultural policy in
Australia.
1.5 Thesis structure
Chapter One has introduced the topic of the thesis and the research questions of the study.
Definitions of HL and HLLs have been provided. Background information related to Chinese
immigrants and Chinese language within multicultural Australia has been introduced. This
chapter also poses the researcher’s subjectivity and highlights the significance of the study.
Chapter Two provides a review of relevant literature and identifies the gaps in the literature.
Based upon the literature review, a theoretical framework is developed in Chapter Three.
Bourdieu’s signature notions of habitus, capital, and field are examined in turn to help
31
construct the theoretical framework. The methodology for this research project is presented in
Chapter Four. The combination of quantitative and qualitative approaches within the study is
explicated. The rationale for the research design is explained in detail. Chapters Five and Six
provide data analysis of the initial quantitative phase and the subsequent qualitative phase
respectively. Main findings addressing the research questions are reported. The thesis
concludes in Chapter Seven by discussing important issues emerging from the data analysis,
outlining the unique contributions of the study, addressing limitations of the research,
highlighting future directions for research, and providing an overall conclusion to the thesis.
33
Chapter 2: Literature Review
As articulated in Chapter One, this study addresses two research questions:
RQ1. Is CHL proficiency of young Chinese Australian adults influenced by their
access to various resources, the strength of their ‘Chineseness’, or both?
RQ2. What does CHL mean to young Chinese Australian adults?
In relation to these two questions, the literature review provided in the present chapter
straddles different bodies of scholarship on commitment to CHL learning and identity
construction in CHL learning. Section 2.1 unpacks the notions of ‘motivation’ and
‘investment’ from a social psychological perspective and a poststructuralist perspective
respectively. This section also evaluates how empirical studies make sense of these two
notions in CHL learning contexts. Section 2.2 reviews social psychological and
poststructuralist understandings of identity, with particular regard to CHLLs’ ethnic identity
construction through their CHL learning contexts. Section 2.3 explicates a Bourdieusian
stance and argues the salience of Bourdieu’s notions of ‘capital’ and ‘habitus’ in the current
study. Section 2.4 summarises the literature review, identifies the gaps in the previous
research, and establishes a conceptual direction for this study.
2.1 Commitment to CHL learning: ‘Motivation’ and ‘investment’
The benefits of learning and retaining HLs, both for individuals and for society, have been
well documented (C. Baker, 2003; Cho, 2000; McGinnis, 2005; Tse, 2001), but in
English-dominant countries, loss of HL occurs with the shift from HL to English (C. Baker,
2003; Fishman, 1991; Tse, 2001). In these contexts, CHLLs’ commitment to CHL learning is
34
apt to vary considerably (D. Li & Duff, 2008). CHLLs may study Chinese because their
parents push them to; or maybe it is easier for them to learn Chinese than other foreign
languages; or perhaps they want to communicate within the Chinese-speaking community
(Xing, 2006). In the past two decades, the rapid economic growth in China has extended
these commitments to include personal benefit and future opportunities of employment
(Tsung & Cruickshank, 2011). The following sections present an extensive review of
empirical research, addressing various reasons behind CHLLs’ different levels and forms of
commitment to CHL learning in different situations.
2.1.1 Motivation: A social psychological perspective
Motivation is the activation of goal-oriented commitment promoted by individuals’ specific
reasons and orientations for doing things. Motivation deals with explanations of why people
do the things they do, how long they are willing to sustain the activity, and how hard they
intend to pursue it (Dörnyei & Skehan, 2003). In language learning contexts, motivation is
defined as the directed exertion by which the learner works or strives to learn the Target
Language (TL) due to a desire to do so and the feeling of accomplishment and satisfaction
experienced in the activity (Gardner, 1985). This definition indicates that motivation is the
combination of desire and effort to learn the language and favourable attitudes towards
learning the language.
Social psychological approaches to motivation seek to explain the individual
characteristics that affect language learning, and sometimes how social context influences
these characteristics. The pioneering studies on motivation with classical social psychological
perspectives in Second Language learning by Gardner and Lambert (1959, 1972) identify two
major types of motivation, namely integrative motivation and instrumental motivation. By
35
their account, integrative motivation is characterised by the learner’s positive attitudes
towards the TL group and the desire to learn the TL that helps the learners to integrate into
the TL community. Integrative motivation includes the desire to learn a language in order to
communicate with people who speak that language, the desire to identify closely with the TL
group, and the desire for cultural understanding of the TL community. The concept of
integrative motivation implies that successful language learning depends on a willingness or
desire to be like valued members of the TL (Gardner, 1968). The access of language learners
to the TL community is a function of learners’ integrative motivation (Norton, 1995). In
contrast, instrumental motivation, by Gardner and Lambert’s (1959, 1972) account, refers to
the desire to fulfil certain pragmatic and utilitarian goals to gain some social or economic
rewards through the TL achievement. It refers to a more functional reason for language
learning, such as getting a job or passing an examination. Some cross-sectional studies on
CHLLs’ motivation have yielded mixed results about learners’ reasons for affiliating with
CHL. They are reported as follows.
Guided by Gardner and Lambert’s (1959, 1972) motivation theory, Wen (1997)
investigated the motivation for learning Chinese by US university students from Asian and
Asian-American backgrounds. Seventy-seven students at the beginning and intermediate
levels of Chinese proficiency participated in the study. Ninety-two percent of the participants
had some level of Chinese due to their Chinese heritage. These participants were regarded as
CHLLs. Data were collected by a questionnaire.
The results of multiple regression analysis showed that integrative interest in Chinese
culture and the desire to understand one’s own cultural heritage were the initial motivation
for students to start learning Chinese. This suggested that integrative motivation played an
important role in the preliminary stage of Chinese learning. Students’ presumption that
36
Chinese courses were less demanding than other courses also motivated them to choose
Chinese. This suggested that beginning students were also instrumentally motivated in
learning Chinese to fulfil course requirements. In addition, the results demonstrated that
instrumental motivation contributed to efforts by intermediate learners. Those who were
motivated to exert effort in learning Chinese for certain practical purposes tended to think
that learning outcomes and performance would lead to certain meaningful results or valued
instrumentality.
Similarly, Yang (2003) followed Gardner and Lambert’s (1972) social psychological
motivation model and surveyed 341 college students enrolled in East Asian languages classes,
looking at their language learning motivational orientations. These students were Chinese,
Japanese, and Korean language learners from both heritage and non-heritage backgrounds.
Although some motivational orientations of East Asian language learners were common to
broader Second Language Acquisition or Foreign Language learning situations, integrative
motivation was shown to be more important than instrumental motivation across ethnic
backgrounds. Heritage was the most important variable affecting their integrative
motivational orientations, especially for Korean and Chinese learners. Of the Chinese
learners, 41% claimed Chinese ethnic background. This study implied that Chinese heritage
might function as an integrative motivation in CHL learning.
Yang’s (2003) findings seem inconsistent with Lu and Li’s (2008) study, which
contended that the 120 CHLLs examined were more influenced by instrumental motivation to
pursue Chinese than NHLLs. Interestingly, Wen’s (2011) investigation involving 176 CHLLs
and 141 NHLLs from three US universities observed no significant group differences in
terms of instrumental motivation. The study indicated that all learners valued the usefulness
37
of Chinese language proficiency and studied the language for future opportunities. Although
all learners endorsed instrumentality, instrumental motivation was a significant predictor for
continuing Chinese studies with CHLLs only, suggesting that the decision to continue their
CHL was closely related to perceived usefulness of the language career-wise, and the
perceived importance of the language in the current global economy.
Li (2005) conducted a survey of 695 students at a western Canadian university.
Ninety-two percent of the students were CHLLs. The author did follow-up interviews with 20
CHLLs, five each from Chinese Mainland, Taiwan, and Hong Kong, as well as five other
ethnic Chinese from other locations. The study revealed that most CHLLs were driven by
both integrative and instrumental orientations. No one studied Chinese for purely academic
reasons. They were motivated to learn more about themselves and their ethnic cultures. At the
same time, they hoped to increase their future career opportunities in relation to the growing
Chinese economy. The study also found that CHLLs’ commitment to CHL learning was
shifting. Unlike their childhood CHL learning experiences, which were reportedly
“annoying”, their university CHL learning experiences were reportedly “interesting”,
“rewarding”, and “important”.
Although Gardner’s integrative versus instrumental orientation approach is prevalent in
motivation studies, it is not the only social psychological approach for motivation research.
Comanaru and Noels (2009) examined the motivation of CHLLs in light of
self-determination theory, which contends that learning commitment results from intrinsic
motivation and some forms of extrinsic motivation. One hundred and forty-five participants
were university-level students enrolled in credit-bearing Chinese courses. They differed in
their degrees of exposure to Chinese. They were divided into three groups: (1) Chinese as
38
First Language; (2) English as First Language with at least one Chinese-speaking parent; and
(3) English as First Language with no Chinese-speaking parent. The first two groups were
treated as CHLLs.
Through a survey, the study examined whether CHLLs with different degrees of
exposure to CHL differed in terms of their motivation. The study also investigated the
implications of intrinsic and extrinsic orientations for commitment to CHL learning. The
more CHLLs felt that learning Chinese was personally meaningful and fun, the more they
engaged in the learning process. There were no motivational differences between the first two
groups of CHLLs but they were more motivated than the third group because they felt much
more strongly that Chinese was a central part of themselves, they felt more pressure from
others to learn Chinese, or they had a stronger self-imposed feeling that they ought to learn
the language.
An open-ended question asked participants to describe their motivation for learning
Chinese. Some participants only claimed intrinsic motivation but most participants cited
extrinsic motivation. Several participants indicated multiple considerations for learning
Chinese, showing a mix of intrinsic and extrinsic orientations.
In summary, traditional dichotomies of motivation become blurred and blended among
CHLLs (D. Li & Duff, 2008). Empirical studies have acknowledged the challenges of
distinguishing types of motivation in CHL contexts. Empirical studies have also revealed that
CHLLs’ motivation in CHL learning is not fixed but shifts over time. These findings pointed
out the limitation of Gardner’s and other analogous motivation theories in CHL contexts. The
theories used in the studies attempted to investigate the link between motivation and language
learning within cognitive, individualistic, and reductionist frameworks. This assumes that
motivation is no different than a static physical entity that individual learners either have or
39
do not have, or either have or do not have enough of; and that the more disposed and
motivated learners are to master a language, the more successful they will be in doing so.
These views put the full blame for ineffective language learning outcomes on the learners, as
they are believed to have failed to sustain a necessary level of commitment and drive in the
language learning process. These theories overlook the fact that motivation and social context
are indivisible and that motivation is anchored in individual learners’ reciprocal relationships
and collective practices (Celik, 2007).
Unlike the classical social psychological approaches to motivation, poststructuralist
theorists view commitment to language learning as a co-construction with discursive and
social structures that cannot be simply and easily compartmentalised into one type or another
originating uniquely from or residing within the individual learner (D. Li & Duff, 2008). The
following section will review studies regarding commitment to CHL learning from a
poststructuralist perspective.
2.1.2 Investment: A poststructuralist perspective
‘Poststructuralism’ refers to an intellectual movement that entails new explorations of
differences between cultures, societies, institutions, and individuals (Holmes et al., 2007). It
is an approach to research that embraces seeming contradictions and questions fixed
categories or structures, oppositional binaries, closed systems, and stable truth (Pavlenko,
2002). Block (2007) briefly explains ‘structuralism’ versus ‘poststructuralism’. By his
account, structuralism is defined as the search for universal and invariant laws of humanity
that are operative at all level of human life; in contrast, poststructuralism, the surpassing of
structuralism, is about moving beyond the search for universal and invariant laws of
humanity to more nuanced, multileveled, and ultimately, complicated framings of the world
40
around us. It investigates how social relations between individuals, groups, and communities
impact on the life chances of individuals at a given time and place (Norton, 2000).
In light of feminist poststructuralism, Norton (1995) argues that the classical social
psychological concept of motivation does not pay attention to the complex and dynamic
relationship between the learner and the social world. She criticises this conceptualisation for
viewing learners as static individuals who are inadequately labelled as a constellation of
binaries, including ‘motivated’ versus ‘unmotivated’. In an attempt to critically conceptualise
the relationship between the language learner and the social world and to succinctly identify
the complex identities of language learners, Norton (1995) challenges the static
conceptualisation of motivation in Gardner and Lambert’s (1959, 1972) model and proposes
the concept of ‘investment’ to recast the concept of ‘motivation’.
‘Investment’, it should be noted, was used earlier by Bourdieu (1986, 1989a) to denote
a form of commitment that will produce returns in the form of material and symbolic profits.
Later, Norton (2000) used ‘investment’ to denote learners’ commitment to learning a
language in order to achieve some imagined future self. By her account, language learners
will expect some return on that investment, which is commensurate with the effort expended
on language learning. The metaphor of investment is economic in origin, but differs from the
economic understanding of investment. The economic approach to investment is underpinned
by the rationality of individuals, implying that investment is based on individual decision
making (Coyle, 2002). In contrast, Norton’s feminist poststructuralist approach to investment
emphasises the complexity of individuals. As a result, language learners’ “investment in the
target language may be complex, contradictory, and in a state of flux” (Norton, 2000, p. 11).
The concept of investment in this context was expanded to include material and symbolic
resources rather than being restricted to economic resources. For this reason, learning a
language is investing with the hope of gaining access to a wider range of symbolic resources,
41
such as education, occupation, and friendship, and material resources, such as real estate and
money (Norton, 2000).
Investment is also viewed as closely connected to the ongoing production of language
learners’ multiple and contradictory social identities (Ellis, 1997; Norton, 1995). Investment
is socially constructed in relations of power, changing over time and space, and possibly
coexisting in contradictory ways within an individual language learner. The concept
presupposes that when language learners speak, they are not only exchanging information but
also constantly restructuring their social identities and negotiating the tensions between these
identities and the social world (Norton, 1995).
A distinction should be made between ‘instrumental motivation’ and ‘investment’.
Instrumental motivation is derived from the desire to acquire a tangible reward for learning
the TL. The notion of instrumental motivation presupposes a unitary and fixed language
learner who desires access to material resources that are the privilege of the TL speaker
(Norton, 1994). In contrast, the desire to acquire returns on the investment is more
complicated than a simple tangible reward. The concept of investment necessarily
complicates the ways in which instrumental motivation has been traditionally understood as a
fixed personality trait. The concept conceives of language learners as having a complex
identity, which is multiple, contradictory, and changeable. The notion recognises that learners’
imagined future uses of TL affect their choices of engagement in the language learning
process. The notion also recognises that the relationship between the learner and other
speakers and the relationship between the learner and the social world play significant roles
in the language learning process.
As with poststructuralist perspectives, recasting ‘motivation’ as ‘investment’ offers a
theoretical instrument to examine how individual commitments to TL learning are shaped and
reshaped in and by particular social contexts and how individual learners are rewarded
42
through their commitment to TL learning at a given time and place (Pavlenko, 2002). Some
existing studies illuminate how CHLLs view their investment in CHL learning as access to a
wider range of resources.
Weger-Guntharp (2006) conducted a study in the Chinese language classroom at a
private American university in a major east coast city. Of the 25 undergraduate participants,
eight students were identified as CHLLs. Their ages ranged from 18 to 22. Drawing insights
from Norton’s (1995) concept of investment, this study investigated CHLLs’ commitment to
CHL learning in their first semester of Chinese classes. Biographical profiles of the
participants, an online questionnaire, and an interview consisting of open-ended questions
were used to collect data.
The results of the online questionnaire demonstrated that CHLLs’ commitment to CHL
learning varied, including the intended use of Chinese for future work, attaining advanced
levels of Chinese, and understanding the cultural heritage, indicating that investment in CHL
helped them gain access to a wider range of (potential) resources. The interview results
indicated that investment was changing and shifting across time. In many cases, participants
felt learning Chinese was not “cool” (Weger-Guntharp, 2006, p. 37) because their parents
made them do so when they were in middle school. However, when they were learning
Chinese at the university, they were fond of learning it instead of being pushed to do so by
their parents.
Wong and Xiao (2010) interviewed 64 CHLLs (30 males and 34 females) in US
universities. Thirty-seven were born in the US and 27 migrated with their parents at a young
age. Forty-five spoke a Chinese dialect or dialects as their HL, and 19 spoke Mandarin as
their HL. Norton’s (1995) concept of investment informed the data analysis. Many of the
43
students admitted that learning Chinese was once an unpleasant activity forced on them by
their parents. Later, however, they considered Chinese learning a wise and worthwhile
investment. They considered Chinese to be a prominent currency in the world economy and
contended that Chinese proficiency would enhance their job prospects and favourably
position them in global markets. All the students reported that Chinese learning would give
them a competitive edge in their careers and help them realise their goals. In addition, most
students viewed Chinese learning as a means to foster their connection to Chinese-speaking
networks. Their investment was expected to pay off financially and socially. The findings
confirmed Norton’s (2000) contention that language learning enables learners to access a
wider range of resources as a return on their investment in language learning.
The findings of the reviewed studies indicate that CHLLs’ commitment to CHL
learning can be considered as an investment. Learners were expecting some added value from
their investment in CHL learning. The findings supported Norton’s (1995) argument that
investment in language learning is a means of acquiring greater access to symbolic and
material resources. That is to say, CHLLs’ investment in language made more resources
available and accessible for them. The findings of some empirical studies (Francis, Archer, &
Mau, 2009; A. Hancock, 2006; Lao, 2004; Weger-Guntharp, 2006; J. Zhang, 2009) also
indicated that an investment in CHL learning was an investment in CHLLs’ identities, ethnic
identity in particular. At the same time, learners’ commitments to language learning were
shaped by the range of identities available to the language learners (Pavlenko, 2002). The
mutual relationship between CHL learning and CHLLs’ identity construction will be
discussed in more detail in the next section.
44
2.2 Identity construction in CHL learning
Biologically, individuals are distinguished by their genetic or physical characteristics,
including but not limited to skin colour, gender, facial features, and body size. Socially,
individuals are produced in conditions that determine individual membership in categories
based on class, religion, culture, education, family, and peer group amongst other things.
These biological and social areas are what ‘identity’ is popularly understood to include.
Historically, the identity-language link has attracted a great deal of attention. Johann
Gottfried von Herder (1744-1803), the German philosopher, observed that even the smallest
nations cherish the great deeds of their forefathers in and through their languages, and
concluded that language is the collective treasure of group feeling (Barnard, 1969). Similarly,
Wilhelm von Humboldt (1767-1835), the Prussian philosopher, insisted that language is the
“spiritual exhalation” of a nation (Cowan, 1963, p. 277).
Following these historical assumptions of a necessary language-identity link, the
mutually constitutive relationship between language and identity construction has been
widely recognised in recent literature. Specifically, studies of CHLLs (Chao, 1997; He, 2008)
have demonstrated that learners may study Chinese to search for their ethnic identities. CHL
has an innate relationship with CHLLs’ ethnic identity (Lin, 2004). This immanent
relationship has been examined in CHL contexts from different theoretical perspectives.
2.2.1 Identity: A social psychological perspective
The notion of ‘identity’ has been the object of extensive scholarly treatment in the social
psychological literature in recent decades. Some definitions of identity are based on the
commonalities belonging to the in-group as well as the uniqueness that distinguishes them
from the out-group. For example, identity is defined as a subjective feeling of sameness and
45
continuity that provides individuals with a sense of self and serves as a guide to choices in
key areas of one’s life (Erikson, 1968). Identity is also defined as part of an individual’s
self-concept that derives from knowledge about the individual’s membership of a social
group (or groups) together with the value and emotional significance attached to that
membership (Tajfel, 1981). These definitions imply that identity is a distinctive character
belonging to a given individual, or shared by all members of a particular group (Rummens,
2003). The character shared by the in-group and distinguished from the out-group determines
the recognition of, and affiliation to, a membership. In these ways, identity marks the ways in
which individuals are the same as others or in which the in-group are different from the
out-group. These understandings are attuned to the psychological notion of ‘self-identity’ that
concerns individuals with the state of being a unique person distinct from all others as
reflexively understood by that individual through time (Rummens, 2003). Individuals are
shaped and formed, in a relatively fixed worldview, by sharing common or similar beliefs and
opinions, common or similar emotional attitudes, as well as common or similar behavioural
dispositions.
The social psychological understanding of ethnic identity makes sense when
investigating ethnic groups’ HL loss, maintenance, or development. Ethnic groups must
collectively decide whether to maintain their HL or to let it erode (Giles & Johnson, 1987). In
other words, HL erosion, maintenance, or development depends, to a certain extent, on
decisions being made and strategies being enacted. Following this traditional approach, some
studies investigated the HL and ethnic identity link from a social psychological perspective. A
few empirical studies addressed Chinese immigrants and their descendants in terms of the
relationship between their ethnic identity and their CHL learning. These studies are reported
below.
46
The primary goal of Kiang’s (2008) study was to examine individual differences in
ethnic label choice among young American adults from Chinese backgrounds. Self-reported
ethnic labels were examined among 242 young American adults with Chinese ancestry (age
range = 18-32 years, M = 23.97; 73% female, 27% male). Approximately 30% were
first-generation immigrants, 64% were second-generation or born in the US, 5% were
third-generation, and 1% were fourth-generation.
Associations between ethnic labels and generation, the ethnic diversity of one’s
community and peer group, ethnic identity, and language proficiency were considered by
using a quantitative approach. Internet technology was used to recruit participants and to
administer self-report online questionnaires. Ethnic labels fell under broad categories
whereby 22% reported heritage national labels (e.g., Chinese), 35% added ‘American’ to
their heritage national label (e.g., Chinese American), and 42% reported panethnic-American
labels (e.g., Asian American). Findings indicated that ethnic identity was linked to greater
CHL proficiency and more positive relationships with parents and Asian peers.
Oh and Fuligni (2010) conducted a cross-ethnicity study. This study focused on the
influence of HL proficiency and use on the social development of adolescents from
immigrant backgrounds. The influence of HL proficiency on adolescents’ ethnic identity was
also examined. The analyses described in the study focused on the 414 participants with
either Latin American or Asian backgrounds (mean age = 14.9 years; 49% male, 50% female,
1% did not report gender) who were either immigrants themselves or who had immigrant
parents. Of the 229 Asian American participants, the majority (67%) were of Chinese
descent.
Correlation analysis revealed significant positive associations between the ethnic
identity subscales and HL proficiency. Levels of ethnic identity also varied with language use.
47
HL proficiency seemed to be more important than language use patterns for family
relationships and ethnic identity. It was HL proficiency, not language use patterns, that was
associated with the quality of parent-adolescent relationships and the strength of ethnic
identity. In the case of family relationships, HL proficiency was found to be associated with
the quality of adolescents’ relationships with their parents, but these relationships did not
vary with language use patterns. As for ethnic identity, though HL proficiency and language
use patterns seemed separately associated with the strength of adolescents’ identification with
their ethnic group, when taken together, it appeared that HL proficiency was the stronger
predictor of ethnic identity.
Empirical findings of a few other studies also indicated the role played by ethnic
identity in the process of CHL learning. The sense of relatedness to Chinese family and
community was found to be the most consistent predictor of a self-determined orientation to
CHL learning (Comanaru & Noels, 2009). Examination of CHLLs’ preferences for simplified
or traditional scripts and their attitudes toward different dialects demonstrated connections
between CHLLs’ evolving identities and reasons for the preference of script and dialect (D.
Li, 2005). It indicated the relationship between learning a CHL dialect and sub-cultural
Chinese identity.
In summary, ethnic identity has been commonly understood in social psychological
treatments as one’s alignment or affiliation with, or membership in a particular ethnic group.
It is also a sense of emotional ties that a person has with the group and the meanings of the
ties to the person. The existing literature about CHLLs’ identity focuses heavily on ethnic
identity. Key areas of investigation include CHLLs’ self-labelling, self-identification, and
sense of attachment and belonging (Kiang, 2008; Rosenthal & Feldman, 1992). Some studies
48
have addressed the positive relationship between ethnic identity and CHL proficiency or
usage (Comanaru & Noels, 2009; Feuerverger, 1991; Kiang, 2008; D. Li, 2005; Oh & Fuligni,
2010; Rosenthal & Feldman, 1992). As such, the emphasis is clearly on how individuals
identify themselves, as well as on how related factors such as place of birth, generation,
gender, age, socialisation, and CHL proficiency, help to inform such self-identifications.
Much of this work falls into the social psychological realm (Rummens, 2003).
2.2.2 Identity: A poststructuralist perspective
In contrast to the social psychological perspective, poststructuralism conceptualises identity
as the conscious and unconscious thoughts and emotions of individuals, their sense of
themselves and others, and their ways of understanding their relation to the world (Norton,
1995, 2000). Identity is the unpredictable outcome of a combination of diverse systems in
which people come to terms with who they are in relation to others around them, and is
strongly influenced by the way they view the past, present, and future (Norton, 2000). These
definitions of identity take account of the relationships amongst individuals, others, and the
social world. Firstly, identity explains how individuals perform, interpret, and project their
sense of themselves by the body they look like, by the language they speak, and by the social
behaviour they adopt (Block, 2007). Secondly, identity offers an idea of how individuals
understand others around them and how they relate to others (Woodward, 1997). Thirdly,
identity examines how a person fits in or belongs to the social world and links our social
positions with social situations (Woodward, 2003). Consequently, identity gives individuals a
location in the world and a link to the society that they belong to (Weerakkody, 2010). It also
makes sense of individuals’ relations to the social, cultural, economic, and political changes
taking place in the social world (Woodward, 2003). These are poststructuralist understandings
49
of identity in that they are concerned with how individuals are formed as subjects, how
individuals adopt their subject position, and how individuals experience social variations
across time and space (Holmes et al., 2007). The marking of commonality and difference is
achieved not only symbolically through representational systems but also socially through
inclusion and exclusion of certain groups of people (Woodward, 1997). In what follows, three
defining characteristics of identity of the poststructuralist perspective are discussed.
Firstly, identity is multiple (Norton, 1995). It is produced by, and at the same time
produces, the meaning-making practices of a variety of social sites, such as home, workplace,
school, and community, in which individuals may take up different ‘subject positions’. Some
positions may be different from others. Although a person may be positioned in a particular
way within a given site, the person might not be considered a passive subject to take the
given or marginalised position. Rather, the person might take up new and powerful positions
that do not displace the previous ones (Lo-Philip, 2010). Therefore, individuals have multiple
identities.
Secondly, identity is a site of struggle (Norton, 1995). This conception of identity as a
site of struggle is an extension of the position that identity is multiple. Poststructuralism
insists that the individual is always at the site of conflict and disunity, embracing quite
contradictory modes of identity at different moments (Weedon, 1997). Given identity
construction in a particular system of meanings and values, individuals may find themselves
resisting alternatives. However, when they move out of familiar circles, through education or
politics, they may be exposed to alternative ways of constituting the meaning of their
experience. These situations may well be contradictory, leading to struggle. For example,
individuals sometimes live across social, geographical, and linguistic border crossings.
Living at the intersections of different cultures with different discourses and expectations,
they must negotiate group boundaries when travelling between, or dealing with, different
50
communities (Root, 1992). These are the boundaries to mark inclusion or exclusion,
integration or isolation, alliance or alienation. Crossing the borders, they find that their
identity is destabilised and that they enter a period of struggle to reach equilibrium (Block,
2002). The outcome of the struggle is not to add something new to the old. Nor is it a
half-and-half situation where individuals become half of what they were and half of what
they have been. Identity is constructed through the contradiction of differences. It is about
negotiation at the crossroads of the past, present, and the future. The entire process is
conflictive and ambivalent as opposed to harmonious. In this frame, identity is understood as
a site of struggle.
Thirdly, poststructuralism also argues that identity is socially constructed and produced
through language (Norton, 1995). Identity is produced in a whole range of discursive
practices in which individuals “are constantly organising and reorganising a sense of who
they are and how they relate to the social world” (Norton, 2000, p. 11). Language is the place
where actual forms of social organisation are defined, where social meanings are constituted
and produced, where individual signs acquire meanings, and where identity is constructed in
ways that are socially specific (Weedon, 1997). In contrast to social psychological
understandings of identity, these arguments imply that identity is neither innate nor
genetically determined, but is socially emerging, modified, transformed, and differentiated
through the use of language. From this perspective, identity is open to continuous redefinition
and constant slippage. As individuals acquire language, they learn to give voice and meaning
to their experiences and to understand language according to particular ways of thinking. This
language learning process constitutes and structures consciousness and sense of self, that is,
identity. As Pennycook (2003, p. 528) states, “It is not that people use language varieties
because of who they are, but rather that we perform who we are (among other things) using
varieties of language.” That is to say, it is through the use of language that identity is
51
expressed, enacted, and symbolised. In line with these poststructuralist perspectives, some
empirical studies have examined CHLLs’ identity construction in CHL learning contexts.
These studies are reported below.
Chao (1997) reported the challenges that Chinese Americans had during their CHL
learning. This study is not about Chinese Americans’ identity constructed in their CHL
learning per se. However, the most noteworthy finding emerging from the data was the
change in cultural identification. When younger, the participants did not appreciate the
importance of their Chinese culture. During their teenage years, they reportedly had a strong
desire to be independent, to be integrated into American culture, and to be accepted as part of
American society. Thus, their priority was to seek identity in the English-speaking
community. In contrast, their Chinese identity was suppressed and CHL, in turn, became an
unnecessary burden that required extra effort to learn. After many years of blending into
American society, the participants gradually realised that they could no longer deny a part of
them that was embedded in CHL and Chinese heritage culture. They then looked at the
opportunity to study Chinese on their own terms. The study reflected the multiple,
contradictory, and shifting identities of Chinese Americans. Their “ethno-racial identity” can
be the impetus towards learning their “native tongue” (Chao, 1997, p. 8).
Shin’s (2010) study examined HL experience and identity of adults with mixed
heritages through in-depth and semi-structured autobiographical interviews. Each participant
had an English-speaking white American parent and an HL (Chinese, Japanese, Korean,
Spanish or Vietnamese)-speaking immigrant parent. The findings suggested that HL
proficiency varied widely. Participants reportedly had different degrees of commitment to HL
learning in order to connect with part of their heritage and to define their identity. Shin’s
52
findings indicated that HL was a considerable facet of mixed-heritage identity. Resonating
with a poststructuralist perspective, the study concluded that mixed-heritage identity was
constructed in dynamic, diversified, and open-ended ways, continuously shaped by changing
social contexts.
Shin’s (2010) findings also addressed some participants with Chinese heritage. One
participant said she looked neither Chinese nor American and she did not really know about
herself. This indicated an ambivalent identity. Though parental support for CHL ranged from
“very active support” to “no support”, participants reported a positive attitude towards CHL
learning. They all pointed out the positive relation between CHL learning and Chinese ethnic
identity.
Wong and Xiao (2010) interviewed 64 CHLLs in US universities. All the students
interviewed reported a contingent side of their ethnic selves. They could be Chinese,
Chinese-American, or somewhere in between, depending on the time, location, and context.
Their findings indicated that CHLLs’ identities are flexible formations, variously possessed,
produced, and practiced through CHL learning process. These findings are consonant with
poststructuralist understandings of identity as changeable with shifts of social space and
passage of social time.
Ang (2001) engages with identity studies in an age of globalisation and diaspora. She
focuses on a transnational condition that spatially and temporally sprawls socio-cultural
formations of people and creates imagined communities with blurred and fluctuating
boundaries sustained by real or symbolic ties to the original homeland. The starting point for
her discussion was the experience of her first trip to China. Though of Chinese descent, she
was born in Indonesia and grew up in the Netherlands, before relocating to Australia as an
53
adult. In China, she was different because she could not speak Chinese. In the West, she was
also different because she looked Chinese. A sense of alienation took hold of her. Due to ‘not
speaking Chinese’, she was positioned as a ‘fake’ Chinese by ‘real’ Chinese. That is, she
failed to legitimise her ‘Chineseness’ because ‘not speaking Chinese’ did not give her a
recognised identity as a ‘real’ Chinese. This was the beginning of her engagement with the
predicaments of ‘Chineseness’ in diaspora. She had the difficulty of constructing the
indeterminacy of ‘Chineseness’ as a signifier for identity.
However, ‘not speaking Chinese’ can cease being a problem for overseas Chinese in
diasporic contexts (Ang, 2001). Being transnational; linking the local and the global; the here
and the there; the past, present, and future, diasporic Chinese have the potential to challenge
static, essentialist, and universal conceptions of ‘Chineseness’. Their complex and flexible
positioning between host countries and China keeps a creative tension between current lived
social place and historical cultural history. This tension fills the space in the bipolar
dichotomy of the present and the past with new forms of culture in the collision of the two
(Ang, 2001), and constructs a syncretic identity of ‘Chineseness’ suitable for the current lived
social world. China, the mythic homeland, need not be the absolute norm for ‘Chineseness’
against which all other Chinese diasporic cultures are measured.
In this mixed-up, interdependent, mobile, and volatile world, clinging to a traditional
notion of ethnic identity is ultimately self-defeating (Ang, 2001). ‘Chineseness’ in a
poststructuralist notion is not fixed and pre-given, but constantly renegotiated and
rearticulated. Diasporic ‘Chineseness’ cannot be envisioned in any unified or homogeneous
way. Rather, it is a diverse, heterogeneous, and ultimately precarious hybridity.
To sum up, poststructuralist approaches acknowledge the multiplicity and
open-endedness of identity processes. They see all individuals as users of multiple linguistic
54
resources and as members of complex layers in multiple communities and societies. Transient
language users are able to move between different and complex contexts. With a
poststructuralist perspective, identity is understood as a cover term for “a range of social
personae, including social statuses, roles, positions, relationships, and institutional and other
relevant community identities one may attempt to claim or assign in the course of social life”
(Ochs, 1993, p. 288). Identity consists of multidimensional, contradictory, and ever-changing
images, descriptions, and evaluations of oneself as active and changing subjects in the eyes of
oneself, others, and the society. Identity is socially constructed rather than born, invented
rather than given, always in a process of change rather than at a standstill. In brief, language
learners’ identities are multiple, dynamic, and contradictory (Norton, 1995; Pavlenko, 2002);
socially produced rather than innate or genetically determined (Weedon, 1997).
2.3 Reframing through a Bourdieusian stance
Both social psychological and poststructuralist schools have made important contributions to
the understanding of language learners’ commitment in certain contexts. Motivation, with its
fixed categories, is understood as an individual entity by classical social psychological
scholarship and is identified as one of the core predictors to language learning outcome. As
such, this perspective puts the full responsibility for unsuccessful learning outcomes on the
learners themselves, without taking enough account of social impacts. In contrast, the
poststructuralist notion of investment accounts for language learners’ commitment in relation
to their lived social world. As a return on this investment, language learners expect to gain
access to a wider range of material and symbolic resources. However, the notion of
investment did not conceptualise the nature of these resources or detail the full learning
trajectories of language learners. In this respect, Bourdieu’s notion of capital can complement
55
the notion of investment because investment demands capital and capital has value or adds
value to realise the return on investment. By definition, ‘capital’ here refers to accumulated
resources and has the potential to produce profits and to reproduce itself in an identical or
expanded form (Bourdieu, 1986).
In relation to understanding identity, social psychological scholarship contends that
ethnic identity is an internal trait of HLLs, while poststructuralist scholarship considers it
socially constructed. On the one hand, some dimensions that are biologically embedded in
ethnic identity cannot be made over and some ethnic perceptions are lasting over time and
across space. On the other hand, dispositions and propensities within a particular ethnic group
may shift over time, being shaped by their lived social world. Bourdieu’s notion of habitus
attempts to think through these two opposing truths. ‘Habitus’ here refers to durable
cognitive structures and a dispositional sense of action that direct people to appropriate
responses to given situations (Bourdieu, 1998). It helps to dissolve a plethora of oppositions
between social psychological and poststructuralist schools. Drawing insights from Bourdieu’s
notion of habitus, identity can be understood as a (un)conscious and dynamic sense of self.
Identity is about how individual dispositions are defined, contained, and enabled by the
bodily features that individuals possess, by the actions that they take, and by the languages
that they speak; and it concerns how individuals internalise various forms of external
elements and produce their sense of self in response to the social structures.
In brief, the social psychological approach focuses on micro-level individual
preferences as the sole explanation of individual language behaviours, while the
poststructuralist approach tries to create macro-level socially constructed factors that shape
individual language choices. This section attempts to reconcile the social psychological
inside-out approach and the poststructuralist outside-in approach. The section develops a
Bourdieusian stance to understand CHLLs’ commitment to CHL learning and their ethnic
56
identity construction through CHL learning. By virtue of Bourdieu’s notions of capital and
habitus, the limitations of both the social psychological and poststructuralist approaches can
be addressed.
2.3.1 ‘Motivation’, ‘investment’, and ‘capital’
In the Second Language context, the concepts of ‘motivation’ and ‘investment’ were
proposed by Gardner and Lambert (1959) and Norton (1995) respectively. The two concepts
of motivation and investment are constructs rooted in significantly different research
frameworks of language learning, and therefore function in different contexts and serve
different research purposes. Gardner and Lambert engaged in quantitative investigations of
motivation as an individual trait from a classical social psychological perspective, whereas
Norton looked qualitatively at language learners’ complex social identities and multiple
desires within the social world and was prompted to reconsider and extend the concept
‘motivation’ from a feminist poststructuralist perspective. As such, Norton’s feminist
poststructuralist alternative, ‘investment’, gives purchase on empirical phenomena that
‘motivation’ cannot.
However, the concept of investment has not gone without criticism over the past
decade. Menard-Warwick (2005) suggested that the construct does not adequately consider
both the structural and agentive forces that shape a learner’s language development.
Moreover, Menard-Warwick faulted the notion of investment for not offering an adequate
investigation into learners’ trajectories. Brief touching on learners’ histories is not enough to
examine the sources of learners’ investment (Menard-Warwick, 2005). In addition, language
learners may face social constraints beyond their control that to some extent restrict what they
could invest, despite the resources and agency that learners bring to their language learning
57
(Menard-Warwick, 2005). To complement the notion of investment, the concept of capital
within certain social spaces has found wide purchase in language education scholarship. With
the complement of the concept ‘capital’, ‘investment’ in language learning fits well with the
economic view of investment as “the process of using resources to produce new capital”
(Case & Fair, 1996, p. 33). In other words, these resources increase the value of capital that
the learners invest in a given social context (Celik, 2007; Norton, 2000). In addition,
Bourdieu’s concept of capital conceptualises various resources into different forms of capital
that language learners can invest and access. According to Bourdieu (1996), language
functions as capital in relation to particular markets, where language learners invest in the
language learning actions to construct social reality and negotiate the very conditions under
which their communicative exchanges take place. The concept ‘capital’ and its various forms
will be developed further in Chapter Three. Although empirical studies did not explicitly use
Bourdieu’s concept of capital in CHL contexts, the capital metaphor could make sense in
these studies. Some cross-sectional studies are reviewed below.
A study identified and analysed factors that may contribute to or impede Chinese
American children’s CHL maintenance (J. Zhang, 2009). Interviews with six children
explored how these CHLLs put symbolic, material, and other resources into the enterprise of
language learning based on a cost-benefit assessment. Only three interviews were reported in
the paper. The three CHLLs were either born in the US or moved to live in the US before the
age of 13.
Zhang’s first case was of a 10-year-old boy Frank, who was eager to speak Chinese, a
language that he was proud of. This boy spent time regularly on CHL and Chinese cultural
activities. His Chinese painting was envied and praised by his classmates. His social studies
teacher often praised him for his bilingual skills. He was excited when recalling these
58
experiences. As a return on this investment in CHL and Chinese cultural activities, he had a
reputation as a Chinese painter and speaker that made him stand out in a good way among his
peers. This reputation could be interpreted as ‘symbolic capital’ in Bourdieu’s term.
Zhang’s second case was of a 14-year-old girl Linda. This girl was confident in
speaking but not reading and writing Mandarin. She was a frequent winner in speech contests
at the Chinese community school. Her sister published an article in the local Chinese
newspaper. She was keen to make faster progress in CHL so that someday she could have an
article published too. She enjoyed attending Chinese community school, as she had friends
there whom she had known for nearly 10 years. The returns on Linda’s investment in CHL
included prizes awarded in Chinese language contests and access to her friends’ network. In
Bourdieu’s terms, these profits of CHL learning could be construed as ‘symbolic capital’,
‘institutionalised cultural capital’, and ‘social capital’.
Zhang’s third case was of a 17-year-old boy George. This boy understood Chinese but
usually spoke English. In contrast to Frank and Linda’s CHL proficiency, his Chinese was
“stuttering and disconnected” (J. Zhang, 2009, p. 203). His parents sent him to China every
other year, each time for approximately two months. His oral Mandarin improved with each
visit. Though he did not enjoy learning in the Chinese community school, he did attend
mathematics classes offered by the school. This helped him in his regular school where he
was labelled a “math genius” (J. Zhang, 2009, p. 204) and assigned the task of tutoring his
classmates. George also liked to associate with his Chinese friends because “it’s easier to get
closer to them and it’s easier to establish friendship” (J. Zhang, 2009, p. 204) than it is with
non-Chinese peers. As a return on his parents’ investment of various resources, George won
the reputation of being a “math genius” and gained more knowledge in math. Moreover, his
CHL also established friendships for him. In Bourdieu’s term, these benefits could be
understood as ‘symbolic capital’, ‘embodied cultural capital’, and ‘social capital’.
59
Francis, Archer, and Mau (2009) conducted an interview study to explore 60
British-Chinese pupils’ discursive constructions of the purposes and benefits of CHL learning
in Chinese community schools. The necessity of learning CHL to facilitate communication
with family members at home and relatives in the motherland was a key theme in the data. In
Bourdieu’s term, CHL proficiency helped tighten family ties as ‘social capital’. Many pupils
also described not being able to speak Chinese as a “disgrace”, “embarrassing”, and a reason
for being “ashamed” (Francis et al., 2009, p. 529). In Bourdieu’s sense, lack of CHL
proficiency could be associated with lack of ‘symbolic capital’. Additionally, some pupils
reported that Chinese learning would facilitate future job opportunities in China. These job
opportunities are convertible into ‘economic capital’ in Bourdieu’s term.
Another study conducted in the US investigated parents’ attitudes toward
Chinese-English bilingual education and CHL usage at home (Lao, 2004). Eighty-six parents
were surveyed. The major reasons for parents’ investment in their children’s CHL included
better career opportunities and effective communication with Chinese-speaking family
members and community. In Bourdieu’s term, these are ‘economic capital’ and ‘social capital’
respectively. The study noted a lack of Chinese books held at home by all the surveyed
families. This is consistent with Xiao’s (2008) study that discovered inadequate Chinese
reading materials in Chinese immigrant homes in the US. The print-poor environment at
home was understood to impede further development of children’s CHL. This indicated that
parents’ investment in ‘objectified cultural capital’ might contribute to children’s CHL
learning.
Luo and Wiseman (2000) examined the familial and peer influences on 250 immigrant
Chinese American children’s CHL maintenance. The results indicated that Chinese-speaking
60
peer influence was the most important factor in Chinese American children’s CHL retention.
Parent-child and grandparent-grandchild cohesion also influenced the children’s CHL
retention. These findings indicated that friendship networks and family environments have
impacts on children’s CHL retention. In Bourdieu’s term, ‘social capital’ can contribute to
CHL proficiency.
The studies reported above did not explicitly use Bourdieu’s concept of capital.
However, Bourdieu’s capital metaphor and typology could shed light on these empirical
studies and give answers to these questions: Why do motivated CHLLs nevertheless fail to
acquire CHL commensurate with their goals? Why does their commitment to CHL learning
differ and vary in that they are sometimes motivated, extroverted, and confident, but
sometimes not? Do they change with different reasons or goals for CHL learning at different
ages, and with different degrees of intensity and effort from one context to the next? What are
the returns on their CHL learning?
2.3.2 ‘Ethnic identity’ and ‘habitus’
In HL literature, HL proficiency, ethnic identity, and the relationship between HL and ethnic
identity have been key foci of investigation (Comanaru & Noels, 2009; Feuerverger, 1989,
1991; Kiang, 2008; D. Li, 2005; Oh & Fuligni, 2010; Rosenthal & Feldman, 1992). The
burgeoning of fruitful cross-disciplinary research examining the relationship between HL and
ethnic identity started in the 1980s (Giles & Johnson, 1987). In the literature, HL is
frequently cited as one of the most important and powerful contributors to ethnic identity
construction. HL is considered to have a key influence on ethnic identity for language
minority individuals (Fishman, 1977; Hurtado & Gurin, 1995; Smolicz, 1981). It has been
61
shown to serve as a symbol of ethnic identity and cultural solidarity to remind language
minority groups about their cultural heritage and transmit group feelings (Giles, Bourhis, &
Taylor, 1977). Other studies (Cho, 2000; Cho et al., 1997) reported that HL development
could be an important part of, and contributor to, identity formation, and could help HLLs
retain a strong sense of identity to their own ethnic group. On the other hand, some studies
have documented that a lack of HL proficiency contributes to descending ethnic identity
because language barriers result in intergenerational conflict as children become frustrated
when they are unable to communicate effectively with family members. When family
relationships weaken and parental influence fades, the older generation is reportedly
hampered in its efforts to transmit ethnic values (Hinton, 1999; Wong-Fillmore, 1991).
Concomitantly, ethnic identity fades away. In short, a consequential link between HL and
ethnic identity has been identified in a variety of studies (Bankston & Zhou, 1995; Fishman,
1989; Imbens-Bailey, 1996).
As argued earlier, many examinations of HL, ethnic identity, and their relationship falls
into the social psychological realm, while a smaller number of examinations of these
problems have been undertaken from a poststructuralist perspective. The latter approach
offers the perspectives that have been absent in social psychological literature: Identities of
HLLs are multiple, shifting, contradictory, and socially constructed. However, the
poststructuralist concept of multiple identities without foundational basis has its limitations
(Luke, 2009). The assumption that human identity is wholly malleable and that the body can
be styled to assume an invented identity runs into problems when faced with the durability of
human beings’ internal schemata (Luke, 2009). The body does remember so that human
beings remain in many ways the products of kinship and blood (Luke, 2009). In line with
Luke’s perspectives, this study introduces the concept of habitus, which will be developed
further in Chapter Three. The notion of habitus underpins people’s dispositions rooted in
62
mind and body, as well as acquired characteristics that are the product of social conditions
and may be totally or partially common to people who have been the product of similar social
conditions (Bourdieu, 2005). As such, Bourdieu’s sociological notion of habitus offers a
useful theoretical lens to examine CHLLs’ ethnic identity.
2.4 Chapter summary
HLLs’ commitment to HL learning and the role played by ethnic identity in HL learning have
been understood differently by different theoretical approaches, varying from the traditional
social psychological approach to the poststructuralist approach. Both approaches offer
meaningful perspectives in certain contexts, but are open to critique in other contexts. The
social psychological camp contends that learners’ motivation and ethnic identity are
structured as static individual traits, whereas the poststructuralist camp considers learners’
investment and ethnic identity multiple, dynamic, and socially constructed. To a certain
extent, these two schools stand in opposition to each other in terms of their different
understanding of HLL’s commitment and ethnic identity. Bourdieu’s sociological notions of
capital and habitus could dissolve a plethora of oppositions between these two camps of
scholarship.
The summary of the literature review is outlined in Figure 2.1. The classical social
psychological notion of motivation considers learners’ commitment to language learning as a
dichotomous individual trait. It puts the full blame for unsuccessful learning on the learners
themselves without taking account of the complex social conditions that structure learners’
language practices. The feminist poststructuralist scholarship recasts ‘motivation’ into
‘investment’, which accounts for the complex relationship between the individual learners
and the social world; however, it does not adequately consider the structural and agentive
63
forces that shape learners’ language development. Since investment demands capital, which
has a capacity to produce new capital or reproduce itself as a return on the investment, the
thesis proposes the use of Bourdieu’s notion of capital to complement the notion of
investment. Likewise, ethnic identity is understood as a set of internal static attributes and
dynamic socially constructed dispositions by the social psychological and poststructuralist
schools respectively. As such, these two schools are largely opposed to each other. Some
physical and biological entities, as dimensions within ethnic identity, cannot be made over or
erased, and do stay with the language learners all through their life. Other dimensions of
ethnic identity, such as cultural awareness and values, are inculcated and cultivated through
language learning, and thus are socially constructed. To reconcile these two camps of
scholarship, the thesis proposes the use of Bourdieu’s notion of habitus to help understand
language learners’ ethnic identity as a set of embodied propensities that have the tendency to
internalise the social structures through language learning.
Figure 2.1. Summary of literature review
Social psychological perspective
on CHL learning
Poststructuralist perspective
on CHL Learning
Bourdieusian perspective
on CHL learning
Motivation
Habitus
Ethnic identity
Ethnic identity
Capital
Investment
Complement
Recast Reconcile
64
In summary, there is a dearth of research investigating HLLs’ commitment to, and their
ethnic identity in, HL learning from a Bourdieusian stance; HL research conducted outside
North America has a relatively small body of literature; and CHL in particular, with its
heterogeneity and complexity in its speakers and their associated social, cultural, and
historical ramifications, has not received its due scholarly attention (He, 2008). In this respect,
there is a demand for sociological investigation into the complexities of CHLLs, not only in
North American but also in other diasporic contexts worldwide where CHL has a relatively
significant population of learners and speakers, such as Australia, a complex lived social
world with an idiosyncratic historical and cultural background. By enquiring into CHL for
Chinese Australians from a Bourdieusian sociological perspective, this study offers a
theoretical and empirical analysis to capture the waxing and waning of Chinese Australians’
CHL learning orientations and stances to fill the gaps in the literature.
65
Chapter 3: Theoretical Framework
The current study draws on Bourdieusian theory to build a framework that underpins the
investigation of the research problems. Capital, habitus, and field are the three main “thinking
tools” (Wacquant, 1989, p. 40) of Bourdieu’s sociological approach. They form an
inter-dependent and co-constructed triad, with none of them primary, dominant, or causal
(Thomson, 2008). They need to be taken together to realise the full value of a Bourdieusian
perspective. Formally, Bourdieu (1989a, p. 101) has offered the equation: “[(habitus)
(capital)] + field = practice”, to summarise the conceptually and empirically essential
relationship amongst capital, habitus, and field, which informs individuals’ practices. The
equation implies that manners of being and thinking, routine behaviours, and patterned
sociocultural activities in which individuals engage (practices) result from the relations
between their dispositions (habitus) and their social resources (capital), within the current
state of play of a particular social arena (field). More specifically, Bourdieu (1991) theorises
that people make language choices as a form of social practice according to the amount of
different resources (capital) as well as the dispositions (habitus) that they have within a given
field. In this chapter, these theoretical insights will be developed to inform this study.
Young Chinese Australian adults may have different levels of CHL proficiency because
they may or may not have learnt or chosen to use their CHL. Their choices form part of their
practices in their social lives. Bourdieu’s three thinking tools offer a conceptual mechanism to
discover the deep-rooted reasons behind the CHL practices of young Chinese Australian
adults. Drawing insights from Bourdieu, this chapter constructs a theoretical framework for
the study to investigate the impact of capital and habitus on CHL proficiency of young
Chinese Australian adults, and the meanings that these Chinese Australians attach to CHL.
66
There are four sections in this chapter. The first three sections (Section 3.1, 3.2, and 3.3)
develop Bourdieu’s key concepts of capital, field, and habitus in turn. In order to relate these
three key concepts to the CHL context and the current study, these sections extend Bourdieu’s
notions of ‘linguistic capital’ and ‘culture’ developed within a single language and culture in
mid-20th
- century France into contexts with multiple languages and cultures in the
21st-century societies. Section 3.4 presents a complete theoretical framework by integrating
the three concepts as a theoretical lens on the research questions.
3.1 Capital
Language learners qua investors differ in terms of the quantity and quality of resources they
possess, and their opportunities for successful language learning vary in proportion to the
resources they possess and invest. It is the learners’ investment of different forms of resources
in the TL learning that will lead to TL proficiency. When learners invest in a TL, they hope to
access and acquire previously unattainable resources as a return on their investment
(Bourdieu, 1991; Norton, 2000). This return on the investment may give them access, not
only to material resources, such as capital goods, real estate, and overall financial success, but
also to symbolic and cultural resources, such as educational qualifications, literature, media,
or friendship. To clarify the concept of ‘investment’, it is useful to examine the concept of
‘capital’ because investment demands capital (Bourdieu, 1990) and the aim of investment is
to add value to one’s capital. According to Bourdieu (1986), capital, “in its objectified or
embodied forms, takes time to accumulate” and has “a potential capacity to produce profits
and to reproduce itself in identical or expanded form” (pp. 241-242). For Bourdieu, capital
refers to all goods, or material and symbolic resources, which present themselves as rare and
worth seeking after in a particular social relation located within a system of exchange. As
67
developed by Bourdieu (1986), capital has four interchangeable and convertible forms:
economic, cultural, social, and symbolic.
3.1.1 Economic capital
‘Economic capital’ refers to forms of financial wealth. It can be “immediately and directly
convertible into money and may be institutionalised in the form of property rights” and other
material objects (Bourdieu, 1986, p. 243). It refers to command over economic resources,
such as cash and assets. Economic capital is at the root of all the other forms of capital
(Bourdieu, 1986) in that other forms of capital are based on economic capital but not simply
reducible to it (Bourdieu, 1986; Dovey, 2010). By this is meant that there is convertibility
between economic and other forms of capital in both directions and at different rates of
exchange in different situations.
3.1.2 Cultural capital
‘Cultural capital’ refers to valued cultural property; legitimate knowledge, behaviour, and
modes of thought; and institutionally recognised capacity that characterise different classes
and groups in relation to specific sets of social forms (Bourdieu, 1986). It is about having the
‘right’ cultural resources and assets. The forms of cultural capital include, but are not limited
to, cultural goods, cultural awareness, language competence, behavioural manner, knowledge,
skills, credentials, and qualifications, acquired through education and upbringing (Bourdieu,
1993a) and accumulated over an extended period (Bourdieu, 1986). These forms can exist in
three states: “the objectified state”, “the embodied state”, and “the institutionalised state”
(Bourdieu, 1986, p. 243).
68
In the objectified state, cultural capital usually appears in the material form of cultural
goods that can be concretely displayed such as books, dictionaries, works of art, and musical
and scientific instruments, which are transferable by legal ownership (Bourdieu, 1986). Use
of objectified cultural capital requires specialised cultural abilities. It is not only ownership of,
but also capacity to choose and consume the cultural goods that really counts (Dovey, 2010).
For Bourdieu (1986), objectified cultural capital is only of value when individuals have a
certain amount of embodied cultural capital to consume the cultural goods. Consequently,
these cultural goods are critical when individuals wield power and obtain profits
proportionate to their mastery of the objectified cultural capital, as well as to the requisite
amount of their embodied cultural capital (Bourdieu, 1986).
In the embodied state, cultural capital refers to cultural inclinations, integrated within
the lasting dispositions of mind and body, existing as behavioural patterns, such as accent,
body language, intonation, facial expression, stance, and attitude, which can engender
authority in social situations and act in a way that shows class or manners. It exists “in the
form of schemata of perception and action, principles of vision and division, and mental
structures” (Bourdieu, 1996, p. 5). It is these embodied dispositions that appreciate and
understand objectified cultural goods (Bourdieu, 1986). The accumulation of embodied
cultural capital demands a process of embodiment and incorporation, and a labour of
inculcation and assimilation; it thus costs time, which must be invested personally by the
investor (Bourdieu, 1986). Hence, embodied cultural capital needs accumulation over time
and cannot be transmitted instantaneously.
In the institutionalised state, cultural capital exists as a connection to certain
institutions (e.g. university and professional associations). It takes an object form, such as
qualifications, certificates, educational degrees, academic titles, and credentials that can be
69
institutionally recognised (Bourdieu, 1986, 1991). Institutionalised cultural capital
symbolises a cultural competence and confers on its holder a conventional, constant, legally
guaranteed value with respect to cultural resources (Bourdieu, 1986).
3.1.3 Social capital
‘Social capital’ is the aggregate of the actual or potential resources that accrue to an
individual or a group by virtue of the possession of a durable network of more or less
institutionalised relationships of mutual acquaintance and recognition, such as family, friends,
clubs, schools, communities, and society (Bourdieu, 1986; Bourdieu & Wacquant, 1992). It is
constituted through “contacts and group memberships, which, through the accumulation of
exchanges, obligations, and shared identities, provide actual or potential support and access
to valued resources” (Bourdieu, 1993a, p. 143). In simple words, social capital is the benefits
that one derives from membership in a group. It is the benefits that one draws from social
cohesion and personal investment in a community and any connections within and between
social networks.
According to Bourdieu, there are three key components constructing different
dimensions of social capital: networks, interactions, and recognised values. Firstly, social
capital is a system of social networks inherent within a community that promotes cooperative
behaviour and serves the specific needs of its members. The size and structure of the network
are related to the volume of social capital contained within it. Secondly, social capital exists
in patterns of social relations, or in processes of interactions between individuals and the
social world. Social capital has value and adds value through these interactions. The strength
of the interactions largely decides the value of social capital. Thirdly, social capital is about
connections to and relationships with less, equally, or more powerful others. The amount of
70
social capital is also associated with the social capital of others with whom individuals are in
relation. The greater the recognised social capital of these others, the greater the social capital
accrued. In brief, the volume of social capital ultimately depends on the size and structure of
the network, the strength of the interactions, and the volume of capital of those to whom the
individual is connected (Bourdieu, 1986).
3.1.4 Symbolic capital
‘Symbolic capital’ is the most elusive form of capital to define. In Bourdieu’s early work, it
denoted the aesthetics of the owner and was a form of cultural capital (Dovey, 2010). In later
writings, symbolic capital appeared to be distinct from any definition of cultural capital as an
individually held resource. Symbolic capital was then defined as follows:
Symbolic capital is any property (any form of capital whether physical, economic,
cultural, or social) when it is perceived by social agents endowed with categories of
perception, which cause them to know it and to recognise it, to give it value. For
example, the concept of honour in Mediterranean societies is a typical form of symbolic
capital which exists only through repute, that is, through the representation that others
have of it to the extent that they share a set of beliefs liable to cause them to perceive
and appreciate certain patterns of conduct as honourable and dishonourable. (Bourdieu,
1998, p. 47)
For Bourdieu, symbolic capital refers to “a reputation for competence and an image of
respectability and honourability” (Bourdieu, 1989a, p. 291). It is known and recognised to be
“more or less synonymous with: standing, good name, honour, fame, prestige, and reputation”
71
(Bourdieu, 1993a, p. 37). Specifically, the manifestations of this capital include accumulated
resources and various key markers of standing, such as prestige, reputation, fame, title,
celebrity, and honour that are founded on the recognition granted by others. This capital
refers to the form assumed by different kinds of capital when they are perceived and
recognised as legitimate (Bourdieu, 1985b, 1989b, 1991, 2000).
Capital, in whatever form, may be convertible into or institutionalised in a title of
nobility (Bourdieu, 1986), existing and acting as symbolic capital (Bourdieu, 2000). As
Bourdieu (1998) further explains:
Symbolic capital is the form taken by any species of capital whenever it is perceived
through categories of perception that are the product of the embodiment of divisions or
of oppositions inscribed in the structure of the distribution of this species of capital. (p.
47)
In this sense, symbolic capital is nothing more than other forms of capital when these are
known and recognised by others (Bourdieu, 1985b, 1989b). Capital, converted into symbolic
capital from whatever form, can be reconvertible back into that capital itself (Bourdieu,
1977b). It is the exhibition of symbolic capital that is one of the mechanisms to make “capital
go to capital” (Bourdieu, 1977b, p. 181). For Bourdieu, symbolic capital acts as a social
mechanism within a system of exchange, specifically, exchange of value and forms among
capital. Symbolic capital can give added value and recognition to all other forms of capital.
As such, symbolic capital is perhaps the most valuable form of accumulation in a society
(Bourdieu, 1977b) because holders of large amounts of symbolic capital are in the position to
impose the scale of values most favourable to their products (Bourdieu, 1989b). Therefore,
72
symbolic capital is the power granted to those who have obtained sufficient recognition to be
in a position to impose recognition (Bourdieu, 1989b). It is subjective and is perceived as
making legitimate demands for recognition (Bourdieu, 1990).
3.1.5 Capitalising on language: Linguistic capital
The concept of capital captures the relevance of the economic metaphor of investment.
Language learners’ investment in a TL demands capital, and in return, these learners hope to
acquire a wider range of resources in the forms of economic, cultural, social, and symbolic
capital through language learning (Norton, 1995). That is to say, the invested capital produces
TL competence and TL competence can reproduce capital when it is valued in a given
situation. Norton’s (1995) interpretation of ‘investment’ resonates with Bourdieu’s (1986)
conceptualisation of ‘capital’, which has the capacity to produce profits and to reproduce
itself in an identical or expanded form. The reproduction of these resources depends on the
value of the profits that this reproduction can expect from its investments of the amount and
the structure of capital in different forms (Bourdieu, 1996). In line with this Bourdieusian
perspective, language competence can be conceptualised as a form of capital because it is
both the outcome of the investment in language learning and the generator of new capital
through the process of language learning. In Bourdieu’s term, it is linguistic capital, and can
be understood as a form of cultural capital (Bourdieu, 1999a), more specifically, as embodied
cultural capital (Bourdieu, 1977a).
Linguistic capital has two dimensions: the mastery of, and relation to, language
(Bourdieu & Passeron, 1990). The mastery of a language can produce linguistic competence
in that language, while the relation to a language can produce legitimate competence in that
language. Linguistic competence is the capacity for “infinite generation of grammatically
73
regular discourse”, which may not be valued either “in the social conditions of its constitution”
or “in the social conditions of its operation” (Bourdieu, 1977a, p. 646). In contrast, legitimate
competence is the capacity to produce judicious and appropriate language, which is used in
strategies invested with all possible functions rather than only communicative functions. To
clarify, what is often rare in practice is not the competence to speak, which, being part of our
biological capacity, is “universal and therefore essentially non-distinctive” (Bourdieu, 1991, p.
55), but rather the competence to speak the legitimate language, which produces “a profit of
distinction on the occasion of each social exchange” (Bourdieu, 1991, p. 55). In this way, in
the mid-20th
French context, Bourdieu differentiated between a literary orientation, which
may include Latinate vocabulary and constructions, as well as a striving for rare and novel
expressions, and a situational orientation, which may include vocabulary that is acquired
rather than learned and reliance upon figures of speech shared by a particular group. In short,
the concept of linguistic capital makes language competence move “from syntax to semantics
and pragmatics” (Bourdieu, 1977a, p. 646). Valued vocabulary, accent, and parlance
(Bourdieu, 1991, 1999b), together with legitimate linguistic strategies, such as “tension or
relaxation”, as well as “vigilance or condescension” (Bourdieu, 1977a, p. 654), can be
valuable aspects of linguistic capital.
Linguistic capital can reflect class-linked traits of speech (Bourdieu & Passeron, 1990).
More generally, linguistic capital reflects the power relationship between a speaker and a
listener. A speaker has to be not only understood but also believed, obeyed, respected, or
distinguished. A listener has to not only understand but also entitle the speaker to speak and
empower the speaker with the right to the legitimate language. As Bourdieu (1977a, p. 648)
explained, “those who speak regard those who listen as worthy to listen and those who listen
regard those who speak as worthy to speak.”
74
Language competence becomes linguistic capital only when it is valued and recognised
in a specific language market. For Bourdieu, there are “hierarchies of legitimacy” or “the
hierarchy of the arts, of genres etc.” (Bourdieu, 1989a, p. 86). As such, there is a hierarchical
situation in terms of the legitimised languages and their value in the linguistic market. That is
to say, different ways of speaking in a language can accrue different amount of linguistic
capital when being legitimised with different values in different situations. Along the
spectrum of a given language competence, there is a hierarchy of different ways of speaking
with different values, from the most legitimised to the least legitimised.
Bourdieu’s concept of linguistic capital also sheds light on people’s choice and pursuit
of different languages in a multicultural society where there is a spectrum of different
languages existing with different legitimised values. This is particularly true for HLs in a
dominant language context. Gogolin (2002) argued that German was positioned at the top of
the hierarchy in the German school education system, followed by different “layers” (p. 126)
of HLs, such as English, Turkish, HLs spoken by legal immigrants and their descendents, and
HLs spoken by illegal immigrants and their descendents. Likewise, there seems to be a
perceived hierarchy among different varieties of Chinese language. In Li and Zhu’s (2011)
study, British Chinese pupils, their parents, and their teachers considered Mandarin most
popular, Cantonese better than Hakka or Hokkien, and other regional varieties “rough” and
“uncultured” (p. 18). Given its potential to illuminate multilingual settings, Bourdieu’s
concept of ‘linguistic capital’ offers a meaningful theoretical tool for the current study to
examine CHL for Chinese immigrants and their descendents in Australia, where linguistic
plurality amongst settlers has been a reality since the first cohort of white immigrants landed
on this continent in the late 18th
century. When CHL is recognised and valued in a given
situation, Chinese Australians may invest their available capital to cultivate their CHL. At the
75
same time, they may expect a return on their investment, an access to a wider range of
different forms of capital. In contrast, Chinese Australians may not engage in CHL learning
when and where their CHL does not accrue any recognised value.
3.1.6 Section summary
As Bourdieu (1986) argues, capital, in its different forms, has the potential to produce profits,
which determine the chances of success for certain practices, and to reproduce itself in an
identical or expanded form through successful practices. Accordingly, linguistic capital can
realise the exchange of value and form through successful language learning. Drawing
insights from Bourdieu’s key concept of capital, the first part of the theoretical framework of
the current study can be modelled (see Figure 3.1).
Figure 3.1. Forms of capital and language
As shown in Figure 3.1, capital has four different forms, namely economic, cultural,
social, and symbolic capital, all of which play a salient role in the process of language
cultivation, a form of practice in Bourdieu’s term. The opportunities for language learners to
cultivate a TL are determined by the quantity and quality of various forms of capital available
to them. When the TL accrues value and becomes linguistic capital in the process of language
cultivation, capitalisation on the TL helps language learners gain access to a wider range of
Capital
Symbolic
capital
Social
capital
Cultural
capital
Economic
capital
Language as
linguistic capital
76
resources in the forms of economic, cultural, social, and symbolic capital. In this respect,
language learning and investment of various forms of capital are related reciprocally. The
reciprocal relationship between capital and language learning is demonstrated by the
double-headed arrow in Figure 3.1.
The reciprocal relationship between capital and language learning only makes sense at
a certain place and time. On the one hand, language learning can be understood as a form of
practice. As Bourdieu (1986) argues, the chances of success for practices are determined by
the structure and the distribution of different forms of capital at a given moment in a given
social situation. According to this perspective, the language market for the investment of the
amount and forms of capital may decide whether a given individual will choose to cultivate a
given language (Garnham & Williams, 1993). On the other hand, the language market
decides the value of the capitalised language, or linguistic capital, and decides the amount
and forms of other capital that this linguistic capital can produce as a return on the investment.
To understand this ‘reciprocal relationship’, it is useful to unpack Bourdieu’s notion ‘field’.
3.2 Field
As argued earlier, all forms of capital are necessarily interrelated because capital is
potentially convertible from one form into another in a given situation. Hence, the concept of
‘field’ is central to understanding Bourdieu’s uses of ‘capital’ in that capital does not exist and
cannot function except in relation to a field (Bourdieu & Wacquant, 1992). A classic
definition of ‘field’ was offered by Bourdieu (2011):
A field is a structured social space, a field of forces, a force field. It contains people
who dominate and people who are dominated. Constant, permanent relationships of
77
inequality operate inside this space, which at the same time becomes a space in which
various actors struggle for the transformation or preservation of the field. All the
individuals in this universe bring to the competition all the relative power at their
disposal. It is this power that defines their position in the field and, as a result, their
strategies. (pp. 40-41)
One of the essential aspects of field is Bourdieu’s view of individuals, groups, and
institutions as agents. As agents, human beings and institutions act consciously with
intentions. By virtue of agents’ consciousness and intention, a field becomes a configuration
of positions comprising agents, who struggle to defend or improve their positions (Bourdieu,
1993b) according to a specific interest of the field, or “practical mastery of its rules”
(Bourdieu & Wacquant, 1992, p. 117). In other words, the position of each particular agent in
a given field is a result of interaction within the specific rules of the field. The positioning
rules in fields are associated with forces acting on the resources at stake. Agents struggle and
compete to take up and occupy these positions in fields according to how many different
forms of symbolic and material resources they have (Bourdieu, 1993b; Bourdieu & Wacquant,
1992; Kramsch, 2008). The field structures strategic action for control over resources that are
construed as forms of capital (Bourdieu, 1991, 1993b). That is to say, to enter a field requires
capital and that capital in turn, enables positions in the field (Bourdieu, 1986, 1996; Lingard,
Rawolle, & Taylor, 2005). Agents within the field compete for control over the capital valued
by the field, and accumulate, conserve, convert, or utilise their different forms of capital in
this competition in order to secure a position within the field. Agents are positioned within a
given field in the first instance according to the overall amount and value of the capital they
possess; in the second instance, they are distributed according to the composition of their
78
capital, that is, the relative weight of different forms of capital in their total set of assets
(Bourdieu, 1985b, 1989b, 1991). Therefore, a field contains a network or a configuration of
relations between positions imposed upon agents by a) their present and potential distribution
of forms of capital, the possession of which commands access to specific profits that are at
stake in the field; and b) their relations to other positions, which could be dominant,
subordinate, or equivalent according to the rules of the field (Wacquant, 1989). In brief, a
field is a structured social space of relational, intentional, and multidimensional activities,
interactions, and transactions that contain agents with different positions according to the
rules of the field.
Not only does capital enable agents to enter a field and place them in a favourable
position in that field, but also capital grants agents ‘symbolic power’ to influence the rules of
the field. According to Bourdieu (1989b), symbolic power is the power to preserve or to
transform objective principles of union and separation, of association and dissociation; and to
conserve or to transform current classifications in matters of gender, nation, region, age,
social status, and language. Agents who capture more capital in various forms can wield their
symbolic power and influence the rules of the field according to their own interest. To this
end, they impose ‘symbolic violence’ on the less powerful others. According to Bourdieu
(1989a), symbolic violence is fundamentally the imposition of thoughts and perceptions upon
the dominated agents that tends to perpetuate the structures of the legitimated actions of the
dominant agents. As such, symbolic violence generates power relations that require the
dominant to dominate and the dominated to accept the domination.
All capital does not necessarily have the same value in a given field because different
fields with which the agents come into contact do not operate under the same rules; capital
only exists and only produces its effects in the field in which it is produced and reproduced
79
(Bourdieu, 1989a). That is to say, each field only generates and consolidates the values
recognised in the field and rewards the relevant capital by adding value to it. It is the field
that produces, reproduces, and legitimises what counts as resources and how they are
allocated. A given field may place a higher value on one form of capital than on another,
although that form of capital may be worth less in another field. The existence, efficacy, and
value of capital depend upon the nature of the rules within the field (Bourdieu, 1993b) in
such a way that capital is produced differently in different fields (Bourdieu, 1996). As
Bourdieu and Wacquant (1992, p. 117) further explain, “each field simultaneously
presupposes and generates a specific form of interest incommensurable with those that have
currency elsewhere.”
A field does not have a fixed boundary. It is dependent on the existence of other fields.
Consequently, a field lies along a continuum between autonomy and heteronomy (Bourdieu,
2000). Each field has enjoyed a certain degree of autonomy, defined by its ability to reject
external interventions and obey its own specific logic, interest, and rules (Bourdieu, 1993a).
When agents operate in a given field, they have to follow the internal rules of that field.
Agents without certain forms of capital valued by the field can be denied entrance into that
field. This demonstrates the autonomy of fields. While fields are relatively autonomous, they
are also constantly (re)shaped by both internal struggles and external developments in related
fields (Jackson, 2009). Internally, various agents struggle for transformation or preservation
of the field (Bourdieu, 1998). In other words, fields do not exist independently of the
consciousness of the agents; rather, fields are (re)produced by agents’ intentions. Externally,
every field may be part of one or several other larger fields, or may contain subfields itself.
No field enjoys complete autonomy or exists independently of other fields. In brief, a given
field imposes on, and is imposed on by both internal dynamics and external interventions.
80
In line with Bourdieu’s theorisation of field, the one degree of language proficiency
may be linguistically equal but not socially equal across different fields, in other words, may
be worth more or less as linguistic capital. In a social field where multiple languages co-exist,
when one language dominates the linguistic community, the linguistic field will make that
language the norm against which the value of other languages is measured (Bourdieu, 1977a).
This recognised domination will, in return, structure the rules of the field to measure other
languages against the legitimate language. Given the essential interdependent relationship
between capital and field, the theoretical framework demonstrated in Figure 3.1 can now be
elaborated into Figure 3.2 (Note the addition of the concept of field).
Figure 3.2. Forms of capital and language in fields
Although there is no research conceptualising Bourdieu’s ‘capital’ and ‘field’ in CHL
contexts, some empirical studies have implied the mutual constitution between ‘capital’ and
‘field’. For example, CHLLs may use CHL in personal conversations among themselves so
that no one else can understand them (J. Zhang, 2009). Entry into this field of conversation
demands CHL as linguistic capital, without which entry will be rejected. Furthermore, CHL
can be used to favourably position CHLLs in the economic field (Wong & Xiao, 2010). For
Fields
Capital
Symbolic
capital
Social
capital
Cultural
capital
Economic
capital
Language as
linguistic capital
81
example, CHL can be valuable cultural capital and symbolic capital in the global job market,
enabling CHLLs to become members of the dominant economic group. In the context of the
current study, Bourdieu’s notion of field helps explain how the structures within Chinese
Australians’ lived worlds shape their CHL learning and legitimise the value of their CHL
learning.
3.3 Habitus
The concept ‘field’ cannot be understood independently of the concept ‘capital’. Nor can it be
divorced from the concept ‘habitus’ (Bourdieu & Wacquant, 1992). A precursor of Bourdieu’s
concept of habitus can be found in Durkheim’s (1964) conceptualisation of the ‘collective
habit’. For Durkheim, the collective habit exists not only in an immanent state in the
successive acts that it generates, but also in the biological realm where it is permanently
expressed for all in a formula that is repeated from mouth to mouth, transmitted by education,
and fixed in writing. Similarly, Bourdieu (1998) argues that agents are endowed with durable
cognitive structures and a dispositional sense of action that direct them to appropriate
responses to given situations. This is what Bourdieu means by ‘habitus’. The habitus
integrates past experiences and functions at every moment as a matrix of perceptions,
appreciations, and actions, making possible the achievement of infinitely diversified tasks
(Bourdieu, 1977b). It refers to the internalisation of the external (Bourdieu, 1988, 1989a) by
which basic dimensions of external conditions in social life, such as gender, race, ethnicity,
norms, class, and occupation come to shape agents’ internal attitudes, values, perceptions,
and dispositions in ways of which they are seldom aware. This latter point is crucial. The
continuity and regularity of agents’ habitus is not a rational mechanism and action is not
principally a matter of rational choice (Bourdieu, 1977b). It is a generative rather than
82
determining mechanism. It is a mechanism that underpins agents’ being, doing, and thinking;
but does not decide them. This mechanism will now be described in some detail.
3.3.1 Conceptualising habitus
Bourdieu used a variety of wordings to explain what he meant by habitus. The following,
commonly cited definition is used here:
Systems of durable, transposable dispositions, structured structures predisposed to
function as structuring structures, that is, as principles of the generation and structuring
of practices and representations which can be objectively regulated and regular without
in any way being the product of obedience to rules, objectively adapted to their goals
without presupposing a conscious aiming at ends or an express mastery of the
operations necessary in order to attain them and, being all this, collectively orchestrated
without being the product of the orchestrating action of a conductor. (Bourdieu, 1977b,
p. 72)
In this classic definition, the term ‘disposition’ seems particularly suited to express what is
covered by the concept of habitus (Bourdieu, 1977b). A system of dispositions encompasses
several distinct meanings: the “result of an organising action” or a set of outcomes of internal
and external “structure”; a “way of being” or a “habitual state”; and a “predisposition”,
“tendency”, “propensity”, or “inclination” (Bourdieu, 1977b, p. 214). This classic definition
highlights the key features of habitus, namely its durable and transposable dispositions and its
structured and structuring structures (Bourdieu, 1977b). Elaboration of these attributes
follows.
83
Firstly, habitus is durable and transposable but not immutable (Bourdieu & Wacquant,
1992). Habitus is durable in that it lasts over time, and transposable in being capable of
becoming active within a wide variety of fields (Bourdieu, 1993b). This is demonstrated in
the way in which the relationship between ‘field’ and ‘habitus’ functions to produce agents’
deportment or manner. As the product of habitus, bodily dispositions are individual, subject to,
belonging to, and characteristic of the agents themselves. In other words, habitus is embodied
and expressed through dispositions, or durable and transposable ways of “standing, speaking,
walking, and thereby of feeling and thinking” (Bourdieu, 1990, p. 70). Habitus is the bodily
inscription of agents’ present and past positions in the social structure carried by individuals
at all times and in all places (Bourdieu, 1977b). Although habitus is a past that survives in the
present and tends to perpetuate itself into the future (Bourdieu, 1977b), it is not immutable.
Habitus can change constantly in response to new experiences (Bourdieu, 2000). As Bourdieu
and Wacquant (1992, p. 133) argue, “being the product of history, it is an open system of
dispositions that is constantly subjected to experiences, and therefore constantly affected by
them in a way that either reinforces or modifies its structures.” The body is “open to the
world, and therefore exposed to the world, and so capable of being conditioned by the world,
shaped by the material and cultural conditions of existence in which it is placed from the
beginning” (Bourdieu, 2000, p. 134). Dispositions, knowledge, and values are always
potentially subject to modification, rather than being passively consumed or reinscribed. This
can occur when narratives, values, and explanations generated by a habitus no longer make
sense; or when agents use their understanding and feel for the rules of the game as a means of
furthering and improving their own standing and capital within a field, as well as adapting
their dispositions. These dispositions therefore behave in ways that necessarily reproduce the
structures but in a modified form.
84
Secondly, Bourdieu defines habitus as a property of social agents, such as individuals,
groups, or institutions, that comprises a “structured and structuring structure” (Bourdieu,
1994, p. 170). Habitus is structured by agents’ past circumstances, such as cultural history,
family upbringing, and educational experiences. Existing knowledge, or the way agents
understand the world, their beliefs, and values, is always structured through habitus, rather
than merely recorded passively. Agents are disposed towards certain attitudes, values, or
ways of behaving because of the influence exerted by their cultural trajectories. Habitus is
structured by these shared cultural trajectories and the conditions associated with a particular
type of existence. As Durkheim (1938) argued earlier,
In each of us, in varying proportions, there is part of yesterday’s man [sic]; it is
yesterday’s man who inevitably predominates in us, since the present amounts to little
compared with the long past in the course of which we were formed and from which we
result. Yet we do not sense this man of the past, because he is inveterate in us; he makes
up the unconscious part of ourselves. (p. 70)
In this sense, habitus is a structured structure. At the same time, habitus is structuring agents’
present and future practices. It is always structuring the moments of practice when a set of
dispositions meets a particular problem or choice in a given context or social space. In this
regard, habitus generates practices, beliefs, perceptions, and feelings in accordance with its
own structure (Maton, 2008). These practices, beliefs, perceptions, and feelings in return
shape agents’ future possibilities and set agents on a particular path that further shapes their
understanding of themselves and of the world (Maton, 2008). As such, habitus is a structuring
structure.
85
Thirdly, habitus is durable and transposable because it is a structured structure. By this
expression, Bourdieu refers to the largely unconscious acceptance of rules, values, and
dispositions as ‘habitus’. This habitus is largely unconscious because it is structured inside
the agents’ body. It is internalised and preconscious, and will be primarily responsible for
generating many of the actions that agents take. The agents rarely have to think rationally
about how to behave or what to say. Their bodies know. As Bourdieu (1987) further states,
Habitus entertains with the social world which has produced it a real ontological
complicity, the source of cognition without consciousness, of an intentionality without
intention, and a practical mastery of the world’s regularities which allows one to
anticipate in the future without even needing to posit it as such. (p. 12)
In other words, habitus is the source of series of actions by agents “which are objectively
organised as strategies without being the product of a genuine strategic intention” (Bourdieu,
1977b, p. 73). Therefore, the durable and transposable habitus is unconsciously structured
and generative.
Fourthly, habitus is not immutable because it is also a structuring structure. Values,
tendencies, and dispositions allow people to respond to shifting cultural rules and contexts in
a variety of ways because habitus allows for improvisations. Hence, Bourdieu defines habitus
as “the durably installed generative principle of regulated improvisations” (Bourdieu, 1977b,
p. 78) and considers the work of habitus as “the intentionless invention of regulated
improvisation” (Bourdieu, 1977b, p. 79). The ‘regulated improvisation’ suggests that habitus
can be transformed and modified (Pahl, 2008). It is the improvisation that offers the
possibility of change within a field (Pahl, 2008). Every agent, wittingly or unwittingly, is both
86
a producer and a reproducer of social structures. Agents’ responses are always largely shaped
and regulated by their habitus. When agents’ habitus, or natural self, is affirmed by people
around them in conformity with the expected norms, their habitus coincides with that of
others who belong to the same culture, class, and way of life in the situation. Therefore,
habitus is not immutable because it is structuring the consciousness in an ongoing process.
Putting these attributes together, habitus can be thought of as the underpinning
mediation that produces agents’ practices, without either explicit reason or signifying intent,
to be none the less sensible and reasonable (Bourdieu, 1977b). Habitus serves as a cognitive
map that routinely guides and evaluates agents’ choices and options. Habitus operates at a
level that is simultaneously conscious and unconscious. Agents do think and act in strategic
ways, and try to use the rules of the field to their advantage, but at the same time they are
influenced, or almost driven, by the values and expectations that they have internalised in
their habitus. Though they may be conscious of making moves and acting strategically, they
are unaware that their motives, goals, and aspirations are not spontaneous or natural, but are
derived from the habitus. Habitus represents a set of dispositions and taken-for-granted ways
of thinking and behaving that agents gradually develop over time. It is reflective of agents’
day-to-day experiences and the routine set of social relationships within which they are
embedded. As such, habitus can be regarded as the internalisation or embodiment in
individuals of the social environment within which they are located (Connolly, 2011).
Habitus is not a random or unpatterned structure but a structure systematically ordered in the
embodied state, persisting across time as a system of dispositions that generate perceptions,
appreciations, and practices (Bourdieu, 1990).
Of particular relevance to this study, habitus can be understood as values, tendencies,
deportment or manner, and dispositions gained from cultural history and social experience in
87
fields. It makes agents think and act in certain ways that are inculcated and structured by
cultural and social history. Bourdieu (1993b) has explained:
The habitus, as the word implies, is that which one has acquired, but which has become
durably incorporated in the body in the form of permanent dispositions. So the term
constantly reminds us that it refers to something historical, linked to individual history,
and that it belongs to a genetic mode of thought, as opposed to essentialist modes of
thought. (p. 86)
Habitus captures how people carry their culture, experience, and history within themselves,
and how they make choices to act in certain ways rather than others. As a system of
dispositions to certain practices, habitus constructs an objective basis for regular modes of
behaviour (Bourdieu, 1994). These modes of behaviour can be predicted by virtue of the
effect of the habitus because agents who are equipped with it will behave in a certain way in
certain circumstances (Bourdieu, 1994). In these behavioural practices, habitus has the
tendency to reproduce itself. It can produce agents endowed with the systems of dispositions
that are capable of engendering the behavioural practices adapted to the field and thereby
contributing ultimately to the reproduction of the habitus (Bourdieu, 1973).
As argued at the very beginning of this section, habitus cannot be understood
independently of a given field. Bourdieu and Wacquant (1992) explained the relationship
between habitus and field:
On the one side, it is a relation of conditioning: the field structures the habitus, which is
the product of the embodiment of the immanent necessity of a field. On the other side,
88
it is a relation of knowledge or cognitive construction. Habitus contributes to
constituting the field as a meaningful world: a world endowed with sense and value, in
which it is worth investing one’s energy. (p. 127)
The relationship between habitus and field helps make sense of agents’ language choices in
certain situations, which can be understood as the product of a “sub-set of dispositions
acquired in the course of learning to speak in particular contexts (the family, the peer group,
the school, etc.)” (Bourdieu, 1991, p. 17). These socially constructed dispositions imply a
certain capacity to speak and a certain propensity to say given things in a particular situation.
This capacity can be understood as the control over linguistic capital, which involves both the
linguistic capacity to generate an infinite number of grammatically correct utterances, and the
social capacity to use this competence adequately in a given situation. This argument
illuminates the way that habitus is the generative basis for language choices and also suggests
a causal link between habitus and language choices in a given field (Bourdieu, 1991).
At the same time, there is a widespread belief that embodied dispositions were
produced through usage of a shared language (Joseph, 2009). An empirical study involving
interviews with Chinese parents in a region of central Scotland (A. Hancock, 2006) echoed
Joseph’s argument. Acquiring CHL literacy was perceived by the parents to be linked with
the transmission of traditional Chinese cultural beliefs and values. When asked about the
attitude towards children’s learning Chinese literacy, one parent replied: “Poor Chinese, poor
Chinese person” (A. Hancock, 2006, p. 363). What emerged from this investigation was that
CHL literacy was considered to be inextricably bound together with the maintenance and
development of Chinese cultural dispositions. The fact that agents spend formative years
learning language results in their acquiring a habitus that will endure into their adult life
89
(Joseph, 2009). Therefore, habitus is structured or restructured through language and
language is central to habitus (Joseph, 2009).
For Bourdieu, the mutually constitutive effect between habitus and language choice
and practice existed within a given language, namely French. The current study extends this
theoretical perspective to a mutually constitutive effect between habitus and language choices
and practices across different languages. As such, habitus can be used as a theoretical lens to
trace the language development of learners who encounter structural forces while enacting
their agentive will in a field (Costa, 2010). The third part of the theoretical framework can
now be modelled. As shown in Figure 3.3, the double-headed arrow demonstrates the
mutually constitutive effect between habitus and language learning.
Figure 3.3. Habitus and language in fields
3.3.2 Interpreting ‘Chineseness’ through habitus
Ascribed attributes and identities, such as race, ethnicity, gender, and native language, are not
of people’s own choice, and thereby they remain an embodied presence and cannot be erased
or made over. Habitus explains the tendency to perpetuate these attributes and identities
(Bourdieu, 1996). But people may have their own intentions to hide, alter, redesign, or
garnish their identities by certain degrees (Luke, 2009). As such, certain aspects of identity,
although durable and transposable, are not immutable. Therefore, the concept of habitus, as a
foundational basis of kinship and culture, offers an opportunity to describe the tensions
Language as
linguistic capital
Fields
Habitus
90
between multiple positioning and identities (Luke, 2009) and to examine the relationship
between practices and identities (Pahl, 2008).
Though it is impossible for all individuals of the same social group to have had exactly
the same experiences and in exactly the same order, each individual of the same group is
more likely than individuals of another group to have encountered situations common
amongst members of the group. Accordingly, within a particular field, individuals who
occupy similar positions are likely to have similar dispositions and therefore to produce
similar practices (Bourdieu, 1985b). In line with this perspective, Costello (2005) conceives
of identity as a set of largely unintentional dispositions, using the notion of habitus to
explicate how identity affects cognitive style as well as embodied deportment. Similarly,
Holland and colleagues (Holland, Lachicotte, Skinner, & Cain, 1998) emphasise the
production and transformation of habitus, seeing it as a fundamental but not final or given
aspect of identity construction. This emphasis is shared by Bartlett and Holland (2002), who
use the concept of habitus as a way of describing and analysing identity formation in practice.
Rowsell (2008) also emphasises the interplay between habitus and identity practices. In brief,
identity construction is often less than fully conscious (Bucholtz & Hall, 2009) and is often
tied to habitual practice (Bourdieu, 1977b). Identity draws upon and reflects habitus (Zacher,
2008). Habitus, in this way, comes to generate identity (Rowsell, 2008). Therefore,
Bourdieu’s key concept ‘habitus’ offers a theoretical instrument to examine people’s identity.
One of the ways in which attempts have been made to make sense of ethnic identity has
also been through Bourdieu’s notion of habitus (Connolly, 2011). Drawing insights from this
perspective, Connolly (2011) demonstrates how children have already begun to embody and
internalise the cultural dispositions and ethnic awareness of their respective ethnic groups.
These cultural propensities of the ethnic groups may not stem from people’s own conscious
91
choice, and hereby may remain durable and transposable across different times and places in
people’s lives. These embodied dispositions, such as affiliated cultural, experiential, and
historical memories, stay (Luke, 2009; J. Webb, Schirato, & Danaher, 2002). In this respect,
ethnic and racial dimensions are constitutive of habitus (Cockerham & Hinote, 2009;
Diamond, Randolph, & Spillane, 2004; Horvat & Antonio, 1999) and habitus can be shaped
by these ethnic and racial dimensions (McClelland, 1990; Reay, 2004). This argument
augments Bourdieu’s own focus on class in habitus.
Although the potential of employing Bourdieu’s concept of habitus in ethnicity
research is evident, few studies have related ‘habitus’ to an analysis of ethnicity, Asian
ethnicity in particular. Given this, the current study examines Chinese Australians’ ethnic
identity of ‘Chineseness’ with reference to Bourdieu’s concept of habitus, interpreting it as a
set of durable and transposable tendencies to think and act in such a way that has been
inculcated by their Chinese heritage, in which they share the same tastes, behaviours, values,
and way of life. Although this ethnic identity may change constantly in response to new
experiences, the change is “never radical, because it (habitus) works on the basis of the
premises established in the previous state” (Bourdieu, 2000, p. 161). In short, the construction
of an ethnic identity makes sense through habitus (Jenkins, 1992) and derives from the shared
cultural history (Rowsell, 2008).
Culture is one of the most complicated concepts in the English language mainly
because it has come to be used across several distinct intellectual disciplines (Williams, 1988).
For Bourdieu, culture does not only refer to the objectified cultural capital, or cultural goods
belonging to consumers, but also refers to consumers’ embodied competence in the
consumption of these cultural goods. Culture thereby constitutes appropriate and legitimate
tastes for these cultural goods, where taste is “the faculty of perceiving flavours” and “the
92
capacity to discern aesthetic values” (Bourdieu, 1989a, p. 474). It is a “gift of nature” born
from a family and “the product of upbringing and education” (Bourdieu, 1989a, p. 1). Due to
the socially recognised hierarchy of family origin and educational system, taste, “a system of
classificatory schemes” (Bourdieu, 1989a, p. 174) or “an acquired disposition to differentiate
and appreciate” (Bourdieu, 1989a, p. 466), is predisposed to classify people into different
cultural groups. Consequently, people with different tastes are marked to have different
culture, from the “authentic” culture to the “imitation”, from the “true” culture to the
“popularisation”, or from the “high culture” to the “middle-brow culture” (Bourdieu, 1989a,
p. 250). This hierarchy is what Bourdieu meant by “distinction” (Bourdieu, 1989a). In this
sense, to have the authentic, true, and high culture is to “know the best that has been said and
thought in the world” (Arnold, 1999, p. xxxiii).
However, Arnold’s elitist conception of culture is concise but problematic. What is
considered ‘the best’ is not fixed but highly contingent on time and space. While the colonial
age purported European culture to be the most advanced form of civilisation, the two World
Wars shook this complacency. Those wars in concert with subsequent anti-colonialist
movements contributed to changes in attitudes toward culture and how it was understood (K.
Louie, 2008). As Johann Gottfried von Herder (1744-1803), the German philosopher, argued
against the high and dominant position of European culture purported by the colonial age,
“the very thought of a superior European culture is a blatant insult to the majesty of Nature”
(Williams, 1988, p. 89). Within European culture, different tastes may be used to distinguish
culture considered to be of higher value from that considered to be of lower value. However,
these tastes should not give more value to European culture than non-European cultures.
Along the cultural spectrum, there are different cultures of the same value. Within each
culture, there will be a hierarchy of ‘high’ to ‘low’ culture. In Williams’ term, this is culture
93
conceived as “a general process of intellectual, spiritual, and aesthetic development” (1988, p.
90). This study speaks of culture not only in the hierarchical sense of certain social and
economic groups within a particular field, but also in the specific and variable cultures of
different nations and periods across fields. In Williams’ term, this is culture conceived as “a
particular way of life, whether of a people, a period, a group” (1988, p. 90).
Parallel to Bourdieu’s ‘culture’ in the French context, there exists Confucian culture in
the Chinese context, where to have ‘culture’ generally refers to having a certain amount of
formal education, being at a certain level of social and economic status, and perpetuating
such social norms. Confucianism has traditionally valued education, thereby constructing a
hierarchy of social and economic positions, and legitimate social norms (K. Louie, 2008).
Accordingly, there are high or low forms of culture within Confucian thought. Adherence to
this Confucian value system, or ‘taste’ in Bourdieu’s term, can distinguish a strong sense of
‘Chineseness’ from a weak one.
Attributes associated with ‘Chineseness’ result from both intentional and unintentional
learning. As Bourdieu suggested, both types of learning are made possible by habitus
acquired through culture (Bourdieu, 1989a) and produced through history (Bourdieu, 1990).
Because Confucianism is the bedrock, even the definitive core, of Chinese culture (Tan,
2008), it can therefore be understood to constitute a cultural history or “previous state”
(Bourdieu, 2000, p. 161) for ‘Chineseness’. Since Confucianism is deeply rooted in Chinese
societies and highly valued in Chinese social fields, it has become a generative mechanism
behind Chinese people’s thinking, being, and doing. In this respect, “history turned into
nature” (Bourdieu, 1977b, p. 78) because what historically needed to be durable and
transposable through a process of continuous reproduction is now inscribed through social
regulations, forms, and norms.
94
As a set of embodied dispositions, ‘Chineseness’ is understood as both the
consciousness and the unconsciousness of the Confucian way of doing and understanding
things. As habitus, ‘Chineseness’ rooted in the Confucian culture is durable and transposable.
It is durable because the core Confucian values have an enduring impact on Chinese people
today (W. O. Lee, 1996), being the dynamic force that directs Chinese life and generates
forms of Chinese life (Tan, 2008). It is also transposable because Confucian culture can be
carried out and enacted by Chinese people who reside not just in China but also throughout
the world. Chinese people, wherever they are in the world, “represent a general ‘Chineseness’,
deriving from the Confucian heritage itself”, which helps the understanding of their being and
learning (Biggs & Watkins, 1996, p. 269). For the purpose of this thesis, it is helpful to
introduce the core values of Confucianism that underpin this habitus of ‘Chineseness’.
3.3.3 Confucian dispositions
Confucianism developed from the ethics and philosophy of the Chinese philosopher
Confucius (551-478 BC), whose principal concept was to maintain harmony and order and
thus keep society together without the undue exercise of force (Clayre, 1984). In theory,
Confucianism is a complex system of moral, social, political, philosophical, educational,
quasi-religious, and ideological thought that has influenced the culture of China and some
countries in East Asia and Southeast Asia. It plays an important role in Chinese civilisation
and has a deep impact on Chinese society, Chinese education, Chinese culture, and Chinese
people.
The dimensions of Confucianism have accumulated over the past 2,500 years.
Different historical eras have reflected different interpretations of these dimensions, but there
is a core set of values consisting of Wuchang (五常, Five Constants) and Sizi (四字, Four
95
Virtues). During the Western Han (206 BC-8 AD) Dynasty, Zhongshu Dong (179-104 BC), in
his book called 《春秋繁露》 (Rich Dew of Spring and Autumn), described the classical
dimensions of Wuchang (五常, Five Constants), namely Ren (仁, Benevolence), Yi (义,
Righteousness), Li (礼, Ritual or Propriety), Zhi (智, Knowledge or Wisdom), and Xin (信,
Integrity or Trustworthiness). During the Ming (1368-1644) Dynasty, Zhonglin Xu
(1567-1620), in Chapter 20 of his book called 《封神演义》 (The Legend of Deification),
described the classical dimensions of Sizi (四字, Four Virtues), namely Zhong (忠, Loyalty),
Xiao (孝, Filial Piety), Jie (节, Continence), and Yi (义, Righteousness). These values were
elaborated in ancient publications about Confucianism. The most authoritative ones are Sishu
(《四书》, Four Books) and Wujing (《五经》, Five Classics). The Sishu (《四书》, Four Books)
are Daxue (《大学》, Great Learning), Zhongyong (《中庸》, Doctrine of the Mean), Lunyu
(《论语》, Analects), and Mengzi (《孟子》, Mencius). The Wujing (《五经》, Five Classics)
are Shijing (《诗经》, Classic of Poetry), Shangshu (《尚书》, Classic of History), Liji (《礼
记》, Classic of Rites), Zhouyi (《周易》, Classic of Changes), and Chunqiu (《春秋》, Spring
and Autumn Annals). The texts of these ancient publications consist of fragments not written
by Confucius himself but compiled by his disciples and second-generation disciples and
collated over a long period of time (Lai, 2008). A core set of Confucian values consisting of
Wuchang (五常, Five Constants) and Sizi (四字, Four Virtues) elaborated in these ancient
publications will be explained in the following sections.
The Wuchang (五常, Five Constants)
Ren (仁, Benevolence) and Yi (义, Righteousness) are fundamental values within
Confucianism. Philanthropy is the key practice of Ren (仁, Benevolence) and Yi (义,
96
Righteousness), by which people can establish healthy and harmonious relationships.
Confucianism argues that people should love each other and should “老吾老以及人之老,幼
吾幼以及人之幼” (《孟子·梁惠王上》, Mencius). This means people should respect not
only their own parents, but also others’ parents; and should look after not only their own
children, but also others’ children. Confucius also said that “君子成人之美” (《论语·颜渊》,
Analects). This means that a good person has the virtue to oblige others; in other words, a
good person will pursue every effort to help others achieve their goals or meet their demands.
This can explain why Chinese people often feel uncomfortable about refusing requests from
others. Otherwise, it is hard to maintain their face, or a good impression. These dispositions
of ‘Chineseness’ are associated with the values of Ren (仁, Benevolence) and Yi (义,
Righteousness).
Li (礼, Ritual or Propriety) refers to secular ceremonial behaviour including the
propriety or politeness of everyday life. It was codified and treated as a comprehensive
system of norms. Shaping rituals in a way that leads to the health of society and its people is
one purpose behind Confucian philosophy. With Li (礼, Ritual or Propriety), propriety of
behaviour is internalised and exerts its influence before peoples’ actions are taken. People
behave properly because they want to avoid losing face. With Li (礼, Ritual or Propriety),
people also understand and acknowledge each person’s correct social position. Li (礼, Ritual
or Propriety) builds hierarchical relationships through protocols, assigning everyone a place
in society with a proper form of behaviour. It is used to distinguish between the younger and
the older, or between the student and the teacher. Social disorder is considered to stem from
the failure to call things or people by their proper names. The Confucian solution to such
disorder is Zhengming (正名, rectification of terms) (《论语》, Analects), by which Confucius
was concerned with the use of titles to establish normative relationship obligations (《论语》,
97
Analects). This ideal form of social interaction shaped by Zhengming (正名, rectification of
terms), an aspect of Li (礼,Ritual), is associated with the Chinese disposition that people
often address those who are senior in age and/or position by their title plus surname rather
than their first name. This form of social interaction “is not lubricated with the democratic oil
of warmth and first names, but with the oil of respect”, which functions as an effective
lubricant in a hierarchical Confucian culture (Biggs, 1998, p. 730).
Confucianism pays great heed to Zhi (智, Knowledge or Wisdom). Confucius said that
“朝闻道,夕死可矣” (《论语·里仁》, Analects), which means, “If I had mastered the
knowledge this morning, I would not have regretted it even if I had to die this evening.”
Accordingly, most essential is the Confucian belief in human self-perfection pursued as the
highest purpose of life through personal commitment to learning. Influenced by this
Confucian belief, Chinese people consider learning honourable and firmly believe that
education is of paramount importance in people’s life (Biggs, 1998). In order to be at the top
of society, one must be a scholar. Chinese students’ motivation for socioeconomic
advancement through education leads them to study harder (Salili, Chiu, & Lai, 2001). In
Australia, a case study (Zhao & Singh, 2011) demonstrated that Chinese-Australian parents
have higher academic expectations for their children than Anglo-Australian parents; and they
also attach a higher value to the academic aspects of their children’s life. To some extent, the
Chinese-Australian parents’ approach to their children’s academic education can be attributed
to their ‘Chineseness’ in pursuing Zhi (智, Knowledge or Wisdom).
Xin (信, Integrity or Trustworthiness) can be literally understood as credibility, which
is valued not only in Confucianism but also in all cultures. To be trustworthy, people should
do what they say and keep their promises. Like the other dimensions of Wuchang (五常, Five
Constants), Xin (信, Integrity or Trustworthiness) is an embodied disposition inculcated by
98
Confucian heritage; unlike the other dimensions, Xin (信, Integrity or Trustworthiness) has to
be granted by others. People with Xin (信, Integrity or Trustworthiness) are trusted only when
others think these people are worthy of their trust.
The Wuchang (五常, Five Constants), as its name suggests, is the constant belief
system of Confucianism. It has constructed the foundation of Chinese ethics, considered the
root of humanity. Due to this constant Confucian belief system, the Wuchang (五常, Five
Constants) has an enduring impact on Chinese people, embedded in their dispositions. As
such, the Wuchang (五常, Five Constants) can be considered as the underpinnings of the
habitus of ‘Chineseness’.
The Sizi (四字, Four Virtues)
Zhong (忠, Loyalty) emphasises obligations of the ruled to the ruler, the dominated to the
dominant, and the less powerful to the more powerful, but puts less emphasis on the
obligations of the ruler to the ruled. With Zhong (忠, Loyalty), the social hierarchy is
maintained. Zhong (忠, Loyalty) is also an extension of one’s duties to the country, the nation,
the family, and friends.
Xiao (孝, Filial Piety) characterises the respect that children should show to their
parents. Confucius said that “父母在,不远游,游必有方” (《论语·里仁》, Analects), which
means, “When your parents are alive, do not travel too much; if you have to, you must let
your parents know where you are.” Chinese people often stay close to their family members,
especially close to the older generation. This can be attributed to their ‘Chineseness’ to fulfil
their filial piety to the older generation. Mencius (372-289 BC), one of the disciples of
Confucius, said that “不孝有三,无后为大” (《孟子·离娄上》, Mencius), which means,
99
“There are three things that can challenge Filial Piety. Being without children, in particular
without sons to continue the patrilineal line, challenges Filial Piety the most.” This
disposition of ‘Chineseness’ is unique as it relates to gender, because traditional Chinese
genealogical notions of identity exclude women.
Jie (节, Continence) is an important Chinese virtue. In Shangshu (《尚书·大禹谟》,
Classic of History), Confucianism argues that “克俭于家”, which means, “Frugality is
important in family life.” In Zhouyi (《周易·否》, Classic of Changes), Confucianism argues
that “俭德避难”, which means, “The virtue of frugality overcomes difficulty.” In Zuozhuan
(《左传·庄公二十四年》, Zuo’s Commentary), a book written by Qiuming Zuo (556-451 BC)
to interpret Chunqiu (《春秋》, Spring and Autumn Annals) written by Confucius’ disciples,
Confucianism argues that “俭,德之共也”, which means, “Frugality is a common character of
people with virtue.” Traditionally, Chinese people live frugal lives. This disposition of
‘Chineseness’ can be attributed to the value of Jie (节, Continence).
Yi (义, Righteousness), as one of the dimensions in Sizi (四字, Four Virtues), is
equivalent of that in Wuchang (五常, Five Constants) as explained above.
The Sizi (四字, Four Virtues), deeply rooted in the Confucian cultural history, has
served Chinese people as a guide to moral living and shaped their doings in a particularly
Chinese way. Chinese people’s propensities associated with their habitus of ‘Chineseness’ can
be largely attributed to the core values in the Sizi (四字, Four Virtues).
The Liuyi (六艺, Six Skills)
Besides Wuchang (五常, Five Constants) and Sizi (四字, Four Virtues), Confucianism
requires people to grasp Liuyi (六艺, Six Skills), namely Li (礼, Rites), Yue (乐, Music), She
100
(射, Archery), Yu (御, Equestrianism), Shu (书, Calligraphy), and Shu (数, Mathematics).
These six skills were first documented in Zhouli (《周礼》, The Rites of Zhou Dynasty): “养
国子以道,乃教之六艺:一曰五礼、二曰六乐、三曰五射、四曰五御、五曰六书、六曰
九数”, which means, “Men who excel in these six skills were considered to have reached the
state of perfection.” These skills were also documented in 《三字经》 (Three-Character
Classic) (Wang, 1223-1296): “礼乐射,御书数,古六艺”, which means, “The traditional six
skills are rites, music, archery, equestrianism, calligraphy, and mathematics.” These skills
incorporated both military and civil components. This strongly supports the observation that
Chinese students often excel at music, calligraphy, and mathematics. This argument is
consistent with the Australian data. For instance, Chinese Australian students have been
shown to outperform white Australian students in mathematics (Lokan, Ford, & Greenwood,
1997; Mu, 2012; Zhao & Singh, 2011). Besides the pedagogical and psychological
explanations to this better performance, the ‘Chineseness’ associated with the value of Liuyi
(六艺, Six Skills) may serve as a generative mechanism underpinning Chinese students’
relative strength in mathematics (Mu, 2012).
The Liuyi (六艺, Six Skills) focuses on the cultivation of a life of perfect goodness. It
serves as a foundation of Chinese people’s cultural dispositions associated with their habitus
of ‘Chineseness’. In this respect, the Liuyi (六艺, Six Skills) largely explains Chinese
people’s perceptions and behaviours in relation to their cultural tastes.
The crisis and rejuvenation of Confucianism
Confucianism has developed and evolved over the past 2,500 years. However, several
historical events nearly devastated this tradition. Fenshu Kengru (焚书坑儒, Burning of the
101
books and burying of the scholars) refers to a policy and a sequence of movements in the Qin
Dynasty, between the period of 213 BC and 206 BC. Shiji (《史记·卷 006·秦始皇本纪》,
Chapter 6: The Basic Annals of the First Emperor of Qin, in Records of the Grand Historian)
documents:
臣请史官非秦记皆烧之。非博士官所职,天下敢有藏《诗》、《书》、百家语者,悉
诣守、尉杂等烧之。有敢偶语《诗》、《书》者弃市。以古非今者族。吏见知不举
者与其同罪。令下三十日不烧,黥为城旦。
This historical record can be translated as follows: After Emperor Qin (秦始皇, the first
emperor of China) unified China in 221 BC, his Prime Minister Si Li proposed that all
histories in the imperial archives except those written by the Qin historians be burned; that
Shijing (《诗经》, Classic of Poetry) and Shangshu (《尚书》, Classic of History) be collected
by the local authorities for burning; that anyone discussing these two books be executed; that
anyone using history to criticise the present be put to death, along with their families; that
authorities who failed to report cases that came to their attention be equally guilty; and
anyone who had not burned the two books within 30 days of the decree be banished to build
the Great Wall. Politically, Emperor Qin (秦始皇, the first emperor of China) strengthened
his central power and dictatorship by the movement of Fenshu Kengru (焚书坑儒, Burning
of the books and burying of the scholars). However, culturally, he wreaked irreversible
damage by destroying most collections of ancient Chinese books published prior to the Qin
Dynasty. As Chapter 121 of Shiji (《史记·卷 121·儒林列传》, Records of the Grand Historian)
documents, “及至秦之季世,焚诗书,坑术士,六艺从此缺焉。” This can be translated as
follows: By the end of Qin dynasty, documents of Liuyi (六艺, Six Skills) have nearly
102
disappeared due to the burying of the scholars and the burning of Shijing (《诗经》, Classic of
Poetry) and Shangshu (《尚书》, Classic of History).
Another blow to Confucianism was caused by the Cultural Revolution. This
socio-political movement (1966-76) did significant economic, social, and cultural damage to
the People’s Republic of China (Xi & Gao, 2005). One of the stated goals of the Cultural
Revolution was to bring an end to the Four Olds: Old Ideas, Old Cultures, Old Customs, and
Old Habits (Chen, 1966). Old Ideas refers to Confucian thought and the works of Confucius
and his successors (Chen, 1966). Hence, the slogan accompanying this movement was ‘to
criticise Confucius’. This movement was devastating. At the material level, it caused
dramatic damage to Confucian heritage. Architecture was ransacked; literature and the
classics were burned; paintings were torn apart; and antiques were shattered (Xi & Gao,
2005). At the spiritual level, it attempted to sweep away social norms and human virtues
established by Confucianism over the previous 2,500 years. People arbitrarily used violence
to break rules and laws. Without a spiritual home, China became a chaotic society at that
time.
Starting in the early 1990s and continuing into the 21st century, there has been a
massive rebuilding effort to restore the culture that was destroyed or damaged during the
Cultural Revolution (Xi & Gao, 2005). This includes the re-instatement of the value system
of Confucianism. The resurgence of Confucianism continues into the new century. Today,
Confucianism enjoys a robust rejuvenation as “an indispensable cultural force” indicated in
the “culture craze” and “national learning craze” (Deng, 2011, p. 563). Accompanied by a
strong interest in the rediscovery of Chinese tradition and history, China is returning to its
own roots for inspiration (Deng, 2011).
103
Confucianism has survived the Fenshu Kengru (焚书坑儒, Burning of the books and
burying of the scholars) and the Cultural Revolution. Despite several attempts to destroy it,
Confucianism remains an integral part of the psycho-cultural construct of Chinese people
today (Tan, 2008) and the texture of contemporary Chinese life (Clayre, 1984). It has had a
deep and extensive impact on various dimensions of Chinese society, including history,
education, arts, ethics, religion, law, politics, and the military (Q. Y. Zhang, 2009). Two and a
half thousand years after his death, Confucius could still represent an entire way of thinking
and living: Confucianism. Therefore, Confucianism continues to provide the basis of the core
value of Chinese people (Q. Y. Zhang, 2009).
People of Chinese ancestry, such as Chinese Australians, no matter how much capital
they capture or invest in whatever fields, are likely to retain aspects of their ‘Chineseness’ as
habitus, at least to a certain extent. But the dispositions brought to a given field do not
preclude deliberate remaking of the habitus. Therefore, the habitus of ‘Chineseness’ may vary
in anticipation of, and in response to, the positioning that occurs through structural
distinctions and categories of discourse that constitute rules of exchange within fields
(Bourdieu, 1998). Therefore, when examining the interactions between Chinese Australians’
‘Chineseness’ and their CHL practice, their ‘Chineseness’ should be interpreted as habitus
animated within a particular field and interacting with the power of the field.
3.4 Chapter summary
With the concept of habitus, Bourdieu seeks to grasp the generative principles that underlie
agents’ practices in specific cultural contexts and social settings, or ‘fields’. Hence, particular
practices should be understood as not only the product of the habitus but also the product of
these fields, as well as the product of the relations between the habitus and the fields. These
104
specific fields are structured spaces of interrelations where agents’ positions are determined
by the distribution of different kind of resources, or ‘capital’. Agents struggle to maintain or
change the distribution of the forms of capital according to the rules of the fields. Those rules
determine the value of capital and allow one form of capital to be converted into another.
Therefore, habitus, field, and capital are three entangled concepts when interpreting agents’
practices.
Language learning and its usage, as forms of practices, receive their value and make
sense only in their relations to a field (Bourdieu, 1991). Hence, the positions and the
dispositions of language learners within a given field are essential to understanding of their
language choices. Bourdieu (1991) has theorised that people make language choices
according to the amount of different resources (capital) as well as the dispositions (habitus)
that they have with reference to a particular situation (field). For Bourdieu, these language
choices refer to a particular accent, a certain vocabulary, or an appropriate way of speaking in
a given language that would be legitimised within a given situation. This study extends this
Bourdieusian stance to choices between different languages. As such, the choice of a
particular language in a given field is then affected by the quantity and the quality of capital
and habitus that are available to the agents in that field. Possession of certain forms and
combinations of capital and a certain habitus lead to the ability to anticipate the ‘language
game’. This ability determines whether an agent has a “basic minimum of chances in the
game, and therefore power over the game” and hence can invest in the potential profits of that
game (Bourdieu, 2000, p. 220).
In the CHL context, some empirical studies have indicated the essential roles played by
CHLLs’ commitment and their ethnic identity in CHL learning processes. Bourdieu’s
concepts of ‘capital’, ‘habitus’, and ‘field’ will shed light on these empirical findings.
105
However, there is scant research that integrates capital, habitus, and field into the
investigation of CHLLs’ language choices. In light of the Bourdieusian sociological
perspectives presented here, a theoretical framework was constructed to guide the research
reported in this thesis. The theoretical framework is demonstrated in Figure 3.4, which is the
synthesis of Figure 3.1, 3.2, and 3.3.
Figure 3.4. Theoretical framework
This framework talks back to the informing theories and the existing literature, and
guides the examination of the interdependence of Chinese Australians’ ‘Chineseness’ qua
habitus, their various resources as capital, and their CHL proficiency as a form of linguistic
capital that results from their practice of CHL learning within particular fields. Specifically,
the theoretical framework underpins the investigation of the two research questions, which
can be posed in more theoretical terms as follows:
RQ1. Is CHL proficiency of young Chinese Australian adults influenced by their
investment of capital, the strength of their habitus of ‘Chineseness’, or both?
RQ2. How do young Chinese Australian adults understand their CHL learning in
relation to (potential) profits produced by this linguistic capital in given fields?
The next chapter explicates the methodology that underpins the investigation of these two
research questions.
CHL proficiency as
linguistic capital
Four forms of
capital
Chineseness
qua habitus
Fields
107
Chapter 4: Methodology
This chapter explains the methodological pluralism that underpins the research methodology,
details the research design of the current study, and addresses ethical considerations. Section
4.1 presents an introduction to the philosophical position that combines a positivist ontology
and a constructivist epistemology. It talks back to the informing theory and guides the
research design. Section 4.2 argues the importance of mixed methods design, which enhances
the scope and depth of the study and increases the validity and reliability of the findings. This
section also emphasises the theoretical and practical appropriateness of mixed methods to the
study. Section 4.3 delineates the rationale of the quantitative phase and argues its function in
terms of operationalising the theoretical framework to address RQ1: Is CHL proficiency of
young Chinese Australian adults influenced by their investment of capital, the strength of
their habitus of ‘Chineseness’, or both? The design of the pilot study and the main study are
then presented. Section 4.4 describes the design for the qualitative phase, which functions as
a complement to the initial quantitative phase. The qualitative phase addresses RQ2: How do
young Chinese Australian adults understand their CHL learning in relation to (potential)
profits produced by this linguistic capital in given fields? Section 4.5 addresses ethical
considerations in this project. This chapter concludes with a brief summary in Section 4.6.
4.1 Philosophical position
The positivist ontology is arguably the most prevalent research paradigm in most natural and
many social sciences since the 1930s. Although it considers reality to be independent of the
observer (Burrell & Morgan, 1979), it does not necessarily exclude the input of human minds.
As Albert Einstein argued, physical concepts are the creation of human minds and are not
108
uniquely determined by the external world (Von Glasersfeld, 2001). In this sense, human
beings sometimes describe their world of reality according to their paradigmatic lenses,
taking the position that both the external world and their minds contribute to reality. This
position is crucial. It posits that, firstly, there is a real world independent of human beings’
knowledge and consciousness; and secondly, there is a dimension that includes human beings’
socially constructed knowledge about reality and its effects. This philosophical position
indicates the existence of a real world independent of human beings’ perceptions, theories,
and constructions; and concomitantly assumes that understanding of this real world is
inevitably a construction from human beings’ own perspectives and standpoint (Creswell &
Clark, 2011).
The current study draws on the philosophical position that combines the positivist
ontology and the constructivist epistemology because this particular way of thinking
resonates with the Bourdieusian theoretical stance. Bourdieu’s theoretical approach is
intended to bypass or dissolve a plethora of oppositions and antinomies (Bourdieu, 1991). His
theory is a systematic attempt to move beyond a dichotomy between ‘objectivism’ and
‘subjectivism’. By ‘objectivism’ Bourdieu means “an intellectual orientation to the social
world which seeks to construct the objective relations which structure practices and
representations” (Bourdieu, 1991, p. 11), “independent of individual consciousness and wills”
(Bourdieu, 1981, p. 87). Objectivism presupposes a break with subjective experience. It
attempts to elucidate the structures and principles upon which subjective experience is
dependent but which it cannot grasp. In contrast, by ‘subjectivism’ Bourdieu means “an
intellectual orientation to the social world which seeks to grasp the way the world appears to
the individuals who are situated within it” (Bourdieu, 1991, p. 11). Subjectivism presupposes
the possibility of the comprehension of the lived experience that is a form of knowledge
about the social world where subjective lived experience is central to comprehension and
109
knowledge. Traditionally, objectivism and subjectivism have been opposed to each other.
However, “the coexistence of two opposing truths defines the full truth” of the social world
(Bourdieu, 1981, p. 89). This Bourdieusian dualist perspective indicates that neither
objectivism nor subjectivism alone is an adequate intellectual orientation (Bourdieu, 1991).
For Bourdieu, subjectivism concentrates too much on individual experience and perceptions
of the social world and overlooks the power of structures within the social world, while
objectivism refuses to take account of individual capacity, relegating and shackling
individuals to objective relations of social structure. Therefore, Bourdieu chooses the term
“structuralist constructivism” or “constructivist structuralism” to stress the dialectical
articulation of the two moments (objectivism and subjectivism) in his theory (Bourdieu,
1989b, p. 14; Bourdieu & Wacquant, 1992, p. 11).
One of the ways that Bourdieu thinks through objectivism and subjectivism is his
conceptualisation of ‘agency’ and ‘structure’. ‘Agency’ refers to the capacity of individuals to
act independently and to make their own free choices (Barker, 2005). By contrast, ‘structure’
refers to the recurrent patterned arrangements which seem to influence or limit the choices
and opportunities that individuals possess (Barker, 2005). Social structure establishes
objective regularities, laws, mechanisms, and systems that are capable of guiding or
constraining agents’ practices and representations. At the same time, social structure does not
exist independently of agents’ consciousness and wills. Their subjective agency shapes their
worldview and seeks to grasp the way that the social world appears to them. Consequently,
agency presupposes the possibility of the comprehension of agents’ lived experience that is a
form of knowledge about the structures within the social world. This intellectual orientation
produces what agents come to know as the reality of the social world.
Bourdieu’s work on the conceptualisation of capital, habitus, and field attempts to
reconcile agency and structure, rather than conflate the two (Reay, 1995). In the Bourdieusian
110
nexus of capital, habitus, and field, agents are socialised by an evolving set of roles and
relationships in social domains (fields), where their dispositions (habitus) and various
resources (capital) are at stake. Agents’ capital and habitus are valued and re-valued by the
field and their agency is structured by the field. At the same time, various agents struggle for
the transformation or preservation of the structure of the field (Bourdieu, 1998). In other
words, fields are structured and restructured by the agents who operate their habitus and
capital within the fields. Thus, the field is a “field of forces”, which imposes on agents and
agents in turn conserve or transform the structure of the field by their agency (Bourdieu, 1998,
p. 32). Bourdieu theorises the necessary relations between the concepts of capital and habitus
in the existing fields in which capital and habitus are contextualised. This relational concern
constructs a theoretical foundation for the current study to examine how Chinese Australians’
‘Chineseness’ qua habitus and their various resources as capital are related to their CHL
learning in their lived social fields. As agents, Chinese Australians inhabit the social world
and therefore tend to adjust themselves to the structure of the social world. At the same time,
they create their social lives and generate the structure of the social world, so the social world
is continuously changing due to their dynamic agency. There is therefore a mutually
influential relationship where Chinese Australians shape society by virtue of their agency,
which in turn is structured by society.
4.2 Mixed methods research
Mixed methods research includes both qualitative and quantitative approaches in research
design, data production, and data analysis to answer research questions in a single study
(Tashakkori & Teddlie, 2003). There are contrasting views towards mixed methods research.
On the one hand, in the past there was considerable hostility between the quantitative camp
111
and the qualitative camp (Bourdieu, 1999c; Bourdieu & Wacquant, 1992). These two
approaches were viewed from purist perspectives as distinct and incompatible. Advocates of
quantitative approaches criticised qualitative approaches, considering them imprecise,
subjective, and unfit for making predictions. Alternatively, the qualitative side argued that
quantitative approaches are based on a naïve theory of objectivity and cannot describe the
complexity of social reality. In concert with these disputes, the appropriateness of combining
quantitative and qualitative methods within a single study was questioned (Guba & Lincoln,
1994; Tashakkori & Teddlie, 2003). On the other hand, Bourdieu (1999c) advises against the
opposition between quantitative methods and qualitative methods. He goes on to argue that
the traditional opposition conceals the fact that both quantitative and qualitative approaches
are based on social interaction, taking place within the constraints of social structures that
exert effects not only on the interactions amongst participants in a quantitative study, but also
on the interactions between the researcher and the researched in a qualitative study. In line
with this Bourdieusian perspective, an increasing number of scholars have articulated less
rigid views. Although the ‘either qualitative or quantitative’ perspective has been prevalent
for a long time, a mixture of ‘both qualitative and quantitative’ approaches has gained rapid
ground in recent years (Danermark, Ekström, Jakobsen, & Karlsson, 2002; Haralambos &
Holborn, 2004). There is a need to transcend the polarisation of methods in quantitative and
qualitative terms (Danermark et al., 2002; Luke, 2010), particularly in sociological research
(Sadovnik, 2011).
‘Methodological pluralism’ (Danermark et al., 2002) suggests that there is no
‘universal method’. Both quantitative and qualitative methods have their domains and
relevance. A particular method cannot be excluded a priori. The key issue is the ability to
judge the strength and the weakness, respectively, of a method for a given research problem.
Qualitative and quantitative methods are not mutually exclusive but complementary research
112
strategies. There is value in multimethodological approaches and it is profitable to combine
methods in practical work.
Following the route of ‘methodological pluralism’, the current study adopted a mixed
methods design. The choice of this design was governed, on the one hand, by the research
questions; and on the other hand, by what can be learnt from social complexity with the help
of different methods. It will be recalled that RQ1 asks about general relationships, tendencies,
and predictions, while RQ2 asks about individual perceptions and meanings of each
participant. The mixed methods design can gain a broader and deeper understanding of
different levels of research problems and better facilitate the examination of complex
phenomena than would be possible through a single method of research. Therefore, mixed
methods research has gained increasing popularity and appeared more frequently in a wide
range of social science disciplines (Mactavish & Schleien, 2000). The current project takes
account of both theoretical and practical considerations in its mixed methods design.
From a theoretical perspective, the study draws heavily on Bourdieusian theory.
Bourdieu’s logic of research is “inseparably empirical and theoretical” because “one cannot
think well except in and through theoretically constructed empirical cases” (Bourdieu &
Wacquant, 1992, p. 160). The goal of such research is “to grasp the particularity within the
generality and the generality within the particularity” (Bourdieu & Wacquant, 1992, p. 75)
and to uncover “the universal buried deep within the most particular” (Bourdieu & Wacquant,
1992, p. 44). By ‘generality’, Bourdieu refers to the kinds, facts, and patterns sought by
quantitative approaches; by ‘particularity’, Bourdieu means the contextualised experiences of
particular subjects sought by qualitative approaches. This viewpoint demonstrates Bourdieu’s
attempt to think through the split between quantitative and qualitative positions (J. Webb et
al., 2002). Bourdieu has shown the way for a kind of research that manages to combine both a
quantitative and a qualitative methodology (Kramsch, 2008; Thomson, 2008). He argues that
113
“statistics are bound to be abstract” but with the complement of interviews and texts,
profound phenomenon can be revealed “even more clearly” (Bourdieu, 1996, p. 174). In the
operationalisation of theoretical concepts such as cultural capital and social capital, there is a
need for a combination of various methods (Vryonides, 2007). Moreover, a fuller
understanding of habitus and how it works in identity construction would benefit from the use
of both qualitative and quantitative methods (Connolly, 2011). Thus, there is theoretical
support for the current study to adopt a mixed methods design to investigate the research
problems.
From a practical point of view, a combination of qualitative and quantitative research
strategies enhances the depth and scope of the investigation. ‘Numbers’ not only add
precision to ‘words’ but can also justify or falsify them. Simultaneously, ‘words’ can add
meaning to ‘numbers’ and explain them. The research questions addressed by the current
study are multi-faceted and the phenomena investigated by the study are complicated. The
complicated linguistic histories, profiles, and needs of CHLLs, their diverse language
learning and socialisation processes and outcomes, and the benefits and challenges of
developing their CHL proficiency call for a variety of methods, across the quantitative and
qualitative spectrum (He, 2006). These complex interactions might well demand a mix of
methods. Proper application of mixed methods in the study not only addresses this
complexity but also supports stronger scientific inferences than would be the case were either
qualitative or quantitative research tools to be applied in isolation. The mixed methods design
makes good use of the strengths of both quantitative and qualitative approaches, while the
weaknesses of each approach are potentially minimised. Practically, the mixed methods
design enables the study to depict a more complete portrait of Chinese Australians,
embracing how they negotiate their habitus of ‘Chineseness’ and capitalise on various forms
of capital through learning CHL in Australia.
114
An explanatory mixed methods design was adopted in the current study. The design
consisted of first producing quantitative data to test and explain the theoretical framework
and then producing qualitative data to complement and enrich the quantitative results. This
design can enhance confidence in the trustworthiness, credibility, and reliability of the
research findings, and add depth and scope to the project. An explanatory mixed methods
design can be especially helpful when unexpected results arise from the initial quantitative
phase. The subsequent qualitative phase can examine and interpret these results in detail.
4.3 Quantitative phase
Quantitative research is a type of approach that frames what is to be studied, asks specific and
narrow research questions, collects quantifiable data, analyses the data by using statistics, and
conducts the inquiry in an unbiased and objective manner (Creswell, 2008). In line with these
traits, a quantitative study addresses research questions that seek to establish relationships or
variance. In order to generalise this relationship or variance, a quantitative study provides a
description of trends, attitudes, or opinions of a population by investigating a sample of that
population.
In a quantitative study the literature functions to identify key issues related to the
research and document the importance of these issues (Creswell, 2008). It will be recalled
that the existing knowledge in the literature argues the interdependence amongst the ethnic
identity that CHLLs carry, the various forms of resources that they accrue, and the CHL
proficiency that they achieve. In light of Bourdieu’s key concepts of habitus, capital, and field,
the interaction of these dimensions was articulated in the theoretical framework. Quantitative
research can verify or reject this framework deduced from existing knowledge and theories. If
‘habitus’ and ‘capital’ do represent the internal agency operating in social fields and do
115
represent a set of dispositions and positions that are manifest, to some extent, in the
regulation of particular ways of thinking and behaving, these patterns should be discernible
and thus ultimately measurable to some degree. However, there is little documentation of a
comprehensive quantitative interrogation on the interactions amongst the habitus of
‘Chineseness’ of young Chinese Australian adults, various forms of capital accessible to them,
and their CHL learning practices within certain social fields. The initial quantitative phase
may add new knowledge through the analysis of the relatedness of these concepts.
A survey design has been selected. It offers a systematic method for gathering
information from a sample of entities for the purposes of constructing quantitative descriptors
of the attributes of the larger population of which the entities are members (Groves et al.,
2004). It is one of the most commonly used methods in quantitative studies to collect data to
test the theoretical models (Groves et al., 2004). In the current study, quantifiable data were
analysed by using statistical modelling to allow generalisations and claims to be articulated
about the population.
4.3.1 Research questions and hypotheses
In the quantitative phase, the main aim of the survey is to collect data to help understand the
relationship amongst young Chinese Australian adults’ ‘Chineseness’ qua habitus, their
various resources as capital, and their CHL learning as practice. Bourdieu (1989a) argues that
practice results from the relations between one’s dispositions (habitus) and one’s position in a
social world (capital), within the current state of play of that social arena (field); and that
people make choices about languages according to the amount of different resources (capital)
as well as the dispositions (habitus) they have within certain fields (Bourdieu, 1991). This
theoretical perspective informs RQ1: Is CHL proficiency of young Chinese Australian adults
116
influenced by their investment of capital, the strength of their habitus of ‘Chineseness’, or
both? Corresponding to this research question, the null hypotheses (H0) are:
H01. The habitus of ‘Chineseness’ of young Chinese Australian adults has no
statistically significant impact on their CHL proficiency.
H02. The investment of capital of young Chinese Australian adults has no
statistically significant impact on their CHL proficiency.
Deduced from the overall theoretical framework constructed in Chapter Three (see Figure
3.4), Figure 4.1 serves as the first part of the theoretical framework that drives examination of
RQ1 and hypothesis testing.
Figure 4.1. Theoretical framework guiding the quantitative phase
4.3.2 Target population
A population is a group of individuals who share particular characteristics pertinent to
research questions (Groves et al., 2004). As argued in Chapter One, this study identifies
young Chinese Australian adults as the population under investigation. A target population is
a group of individuals with some common defining characteristics that can be identified and
CHL proficiency as
linguistic capital
Four forms of
capital
Chineseness as
habitus
Fields
117
studied (Groves et al., 2004) and is therefore available to the research. This section reports on
the rationale for framing the target population of the initial quantitative phase.
According to the ABS 2011 Census, 94% of Chinese Australians and 97% of Chinese
Australians who spoke Chinese at home lived in the eight capital cities. It can be argued that
an overwhelming proportion of Chinese Australians live in urban areas. As such, the
quantitative phase targeted young Chinese Australian adults in urban Australia.
Identity develops over time, beginning in childhood, through a particularly salient
process during adolescence and young adulthood (Oh & Fuligni, 2010; Phinney & Ong,
2007). Levinson (1978) called this life period the “novice phase” and argued that the
overriding task of this phase is to move into the adult world and build a life structure (pp.
322-323). During this phase, or specifically the age range between 17 and 33, according to
Levinson, young adults experience a considerable amount of change and instability while
sorting through various possibilities in their social lives. In line with these arguments, the
quantitative phase targeted young Chinese Australian adults, ranging in age between 18 and
35. If born outside Australia, they had to have moved to Australia before the age of 13, as
children below 13 are considered less shaped by their learning experiences (Bhatti, 2002). In
this respect, their language learning experiences in Australia would be more salient in their
lives, and they would have been educated primarily in English, with Chinese becoming their
HL. The Australian-born and the non-Australian-born groups can be included within the same
target population frame because they share many linguistic, cultural, and developmental
experiences (J. Zhang, 2009). They are both CHLLs, as people who immigrated at an early
age and were primarily educated in English in their host country, or children born in the
English speaking country but exposed to their immigrant parents’ or ancestors’ native
language in a community, are likely to become HLLs (Cho, 2000; Cho et al., 1997). In short,
the target population of the quantitative phase refers to young Chinese Australian adults in
118
urban Australia, ranging in age between 18 and 35. If born outside Australia, they had to have
moved to Australia before the age of 13.
This target population consists of a diverse group of people in terms of their citizenship
status, the languages that they speak, and their countries of origins. They could hold
Australian citizenship, dual citizenship, or Australian permanent residency. They may or may
not be able to speak a Chinese language. They could be born in the Chinese Mainland, Hong
Kong, Macau, or Taiwan. They could be born in other Asian countries where Chinese is
widely spoken, such as Singapore, Malaysia, Indonesia, or Vietnam. They could also be born
in Australia or any other Western country.
4.3.3 Instrument design and development
In a survey design, a questionnaire is a form of instrument that participants complete and
return to the researcher (Creswell, 2008). It is widely used because it has several advantages
(Haralambos & Holborn, 2004). Firstly, data analysis is time effective. Data produced by
questionnaires can be analysed in an effective way, using computer software packages.
Secondly, questionnaires are familiar to most people and they generally do not make people
apprehensive. Nearly everyone has had some experience completing questionnaires. Finally,
questionnaires are considered less intrusive. The respondents are free to complete the
questionnaire in their own time. Based on the above advantages, a questionnaire was
designed and applied in the initial quantitative phase.
Online questionnaires are becoming increasingly popular in quantitative research
(Groves et al., 2004). This is especially relevant for the current study conducted in urban
areas, given the widespread use and ease of access to the internet. The technology enables
electronic data collection and data entry. E-based data production is less time consuming and
119
increases efficiency. Data entry errors are reduced because data entry occurs automatically
and technologically rather than manually. Moreover, the online questionnaire was considered
cost-effective for the current sole-researcher and unfunded study involving a large sample
size and a scattered population across varied geographic areas. Therefore, an online
questionnaire was conducted in the quantitative phase by using the survey generation
software, Key Survey.
The questionnaire for the survey was structured as follows. A participation statement
that outlined the project and explained ethical considerations was presented at the very
beginning of the questionnaire. In this way, the researcher introduced the proposed research
to the participants and obtained their informed consent for participation. At the end of the
questionnaire, participants were asked to choose whether they would like to participate in an
interview for the subsequent qualitative phase. Participants who were interested were asked
to leave their names, email addresses, or phone numbers for further contact. An
acknowledgement was presented to the participants upon completion of the questionnaire.
There were four sections in the main body of the questionnaire. The first section collected
demographic and background information from the participants. The next three sections
measured participants’ attitudes towards attributes often associated with ‘Chineseness’,
various resources they invested in CHL learning, and their perceptions of their CHL
proficiency respectively. Accordingly, ‘Chineseness’ qua habitus, four forms of capital,
namely economic, cultural, social, and symbolic capital, and CHL proficiency were treated as
variables in the quantitative phase, reflected by their corresponding questionnaire items. An
overview of the questionnaire items will be presented in Chapter Five.
To measure items mapping to these variables, a 7-point Likert-type scale was used. The
major statistical models applied to the data analysis in the main study were Factor Analysis,
Structural Equation Modelling (SEM), and Regression. These models require the variables to
120
be measured at the interval level at least, whilst a Likert-type scale is ordinal in nature
(Creswell, 2008). However, the errors for treating the Likert-type scale results as interval data
are minimal especially when there are more categories or choices in this scale and the data
exhibits a symmetrical distribution (Binder, 1984; Zumbo & Zimmerman, 1993). Therefore, a
7-point Likert-type scale is widely accepted as a proxy interval level of measurement in line
with common practice in educational research (Lehman, 1991; Tabachnick & Fidell, 2007).
When there is an existing questionnaire testing the variables investigated in a given
study, it is good practice to adapt the questionnaire to the current context if the original
questionnaire has proven reliable and valid in certain research contexts. Self-reporting is a
commonly used approach to measure language proficiency in large-scale questionnaire
studies where the direct testing of language proficiency is difficult (Phinney, Romero, Nava,
& Huang, 2001). Moreover, self-reporting measures of language proficiency have been found
to correlate highly with direct measures of language ability (Oh & Fuligni, 2010). Chinen and
Tucker’s (2005) ‘can do’ questionnaire was originally used to gauge the self-reported
language proficiency of Japanese HLLs. After permission from the authors was obtained, this
instrument was modified to meet the requirement of the quantitative phase and to measure
participants’ perceptions of their CHL listening, speaking, reading, and writing skills.
When there is no existing standardised questionnaire specifically designed to plumb the
variables of the study, items to measure these variables need to be designed. This is the case
for Chinese Australians’ ‘Chineseness’ qua habitus and possession and investment of various
resources as capital. Therefore, an instrument to measure these concepts was developed. This
was a challenging process consisting of planning, constructing, evaluating, and validating.
Detailed steps and strategies were undertaken as follows.
The theoretical ideas underpinning each concept need to be translated into items within
the questionnaire, which makes it possible to take measurements relating to those abstract
121
concepts (Danermark et al., 2002). This process is called ‘operationalisation’.
Operationalisation is an important stage in quantitative studies. This involves breaking the
theoretical concept down into various components or dimensions in order to specify exactly
what is to be measured. The relations between concepts, formulated in the theory, are then
studied in the form of quantitative relationships between variables (Danermark et al., 2002).
Chinese Australians’ ‘Chineseness’ qua habitus, their various resources as capital, and their
CHL proficiency were regarded as ‘latent variables’ in the quantitative analysis. A ‘latent
variable’ refers to a variable that is not directly measurable but related to several variables
that can be measured directly (Field, 2009). These directly measurable variables are called
‘indicators’ or ‘measures’ and can be operationalised through corresponding questionnaire
items that provide quantifiable data for measuring each theoretical concept. As required by
SEM, a minimum number of indicators are needed to represent a latent variable. There is no
upper limit in terms of the indicator numbers. But too many indicators make it difficult to fit
a model to data (Bentler, 1990). As a practical matter, there is a preferred minimum number
of indicators, and in most cases, four is the absolute minimum required for model fit testing
with five to seven indicators considered to adequately reflect most latent variables (R. Ho,
2006). A good measurement needs to be accurate and consistent. Otherwise, the findings will
not be valid and reliable. Therefore, operationalisation requires attention to two core
methodological concerns, namely ‘reliability’ and ‘validity’.
‘Reliability’ refers to the stability and consistency of a measurement (Creswell, 2008).
‘Test-retest reliability’ and ‘inter-item reliability’ are two commonly used measures of
reliability (Hair, Black, Babin, Anderson, & Tatham, 2006). ‘Test-retest reliability’ is the
expectation that there will not be different findings each time the measures are used,
assuming that nothing has changed in what is being measured. For practical reasons, the
quantitative phase will not use ‘test-retest reliability’. The second and more commonly used
122
measure of reliability is ‘inter-item reliability’, which evaluates whether a set of indicators
coheres and maps to one underlying latent variable. Often reliability cannot be determined
until the data are collected and analysed. A pilot study can however be used to generate some
top-level indicators of reliability. Since the variables of the quantitative phase were
considered to be measured at the interval level, the pilot study used Cronbach’s alpha
coefficient to test ‘inter-item reliability’ (R. Ho, 2006).
Reliability does not guarantee validity. It is a necessary but not sufficient condition of
validity (Hair et al., 2006). ‘Validity’ here refers to accuracy and whether the
operationalisation is correctly indicating what it is supposed to (Creswell, 2008, 2009). It is
the extent to which the measurement relates to the underlying latent variable; that is, it argues
whether the data make sense, and are meaningful. This enables the researcher to draw good
conclusions from the sample to the population. ‘Face validity’ indicates whether the
questionnaire items seem to ask what they are supposed to and seem to be getting the desired
results (Nardi, 2006). It indicates the extent to which the content of the items is consistent
with the construct definition, based solely on researcher’s judgment, experts’ advice, and
research literature (Hair et al., 2006). As such, face validity concerns the acceptability of the
test items. This is a legitimate, but not very rigorous, way of assessing validity. ‘Construct
validity’ signals how well a set of questionnaire items as indicators are measuring the
complexity of a variable, that is, a concept being studied (Nardi, 2006). It indicates the extent
to which a set of indicators reflects the theoretical constructs specified as latent variables that
those indicators are designed to measure (Hair et al., 2006).
Satisfying validity requires an in-depth review of literature to examine how the concept
is defined and described from a theoretical point of view. To this end, when designing the
measurements of habitus of ‘Chineseness’ and various forms of capital, literature about
Confucianism and Bourdieu’s interpretation of capital and habitus were reviewed. The key
123
dimensions within the theoretical concepts were identified. This careful consideration of the
literature also revealed how the concepts of habitus and capital have been operationalised in
previous empirical studies. A detailed review follows.
Habitus is a nebulous concept. Accordingly, it is difficult to measure in empirical
research (Sullivan, 2002). Despite this challenge, existing studies have attempted to
operationalise it. Habitus can explain how individuals formulate their expectations and beliefs
(Dumais, 2002; McClelland, 1990). Following this perspective, both Dumais (2002) and
McClelland (1990) operationalised habitus with a set of measures of occupational aspirations.
More relevant to the current study, Cockerham and Hinote (2009) suggested that measures of
collectivities can be especially useful to quantify habitus, where collectivities refer to
collections of agents linked together through particular social relationships and shared norms,
values, ideals, and social perspectives (Cockerham, 2005). In particular, religion and ideology
are examples of collective perspectives (Cockerham, 2005). Ho and colleagues (2012) have
made an attempt to compile a pool of items to measure Chinese collective perspectives by
consulting Confucian classics and sayings associated with Confucianism. Cockerham, Ho,
and others have informed the operationalisation of the habitus of ‘Chineseness’. As argued in
Chapter Three, Confucianism can be understood as a Chinese quasi-religious and ideological
system. Key dimensions of Confucian norms, values, ideals, and social perspectives
documented in Confucian classics and popular classical sayings associated with
Confucianism were used as measures to quantify habitus of ‘Chineseness’.
According to Bourdieu (1990), “the distribution of capital in its different kinds among
the individuals” is quantifiable (p. 135). Following this advice, attempts have been made to
quantify capital. Highest personal educational attainment (Bourdieu, 1973, 1989a; Veenstra,
2009), attendance at high cultural events (such as concerts, galleries, and museums) (De
Graaf & De Graaf, 2000; DiMaggio, 1982; Dumais, 2002; Marks, 2009; Sullivan, 2001;
124
Yamamoto & Brinton, 2010), and ownership of cultural or educational objects (Roscigno &
Ainsworth-Darnell, 1999; Sullivan, 2001; Yamamoto & Brinton, 2010) could be used as
indicators of the presence of cultural capital. Participation in different kinds of social
associations and frequency of contacts with acquaintances, friends, and family members
could be used as indicators of presence of social capital (Veenstra, 2009).
A key issue in the achievement of valid and reliable operationalisation is the
construction of a well-written and manageable questionnaire. Frustrations and noncompliance
can result from an unclear, biased, wordy, or poorly designed questionnaire. Several strategies
to minimise these dangers are reported below.
Firstly, not all questions are necessarily applicable to all participants. Filter questions
are asked to avoid forcing participants to provide opinions when they do not have any
relevance. These are also called branching or contingency questions (Nardi, 2006). For
example, before asking the question “except Australia and your country of origin, where else
have you lived and for how long”, a filter question of “have you ever lived in a country other
than Australia and your country of origin” was asked first. Similarly, before asking “how long
have you formally studied Chinese”, “have you ever formally studied Chinese” was first
asked. Likewise, “have you ever visited China” was asked before asking “how many times
have you visited China”. This filtering technique ensured that branching or contingency
questions were answered only by those who responded in a relevant way to a previous item.
Participants were filtered towards questions applicable only to them.
Secondly, sometimes the answers to a question depend on whose opinions are under
scrutiny. To ask whether the respondents think “family members should live close to one
another”, different results could be prompted by asking “I always prefer to live close to my
family members”. Similarly, to ask “people always prefer to say yes to save face”, different
results could be prompted by asking “to save face I always prefer to say yes”. In the above
125
examples, the original wording was more about a general belief, but the second wording put
the spotlight on the participants’ own personal views. For another example, one original item
was “I am very popular among my Chinese peers”. This item directly scrutinised participants’
own opinions about how popular they were and could push those participants who were
popular to say so, even though they could normally feel uncomfortable about saying this.
This item was reworded as “people consider me very popular among my Chinese peers”. This
item scrutinised participants’ popularity through the eyes of others to avoid any personal
discomfort of self-reporting on this.
The third strategy is to avoid ‘double-barrelled’ questions that actually measure two
things within one item. Such questions often contain the word ‘and’. For example, an original
item “I always keep up to date with current Chinese affairs by watching TV, listening to the
radio, reading newspaper, and surfing online” was replaced by the item “I always keep up to
date with current Chinese affairs by watching TV, listening to the radio, reading newspaper,
or surfing online”. The original item was difficult to answer for someone who used only one
of the media mentioned not the rest. For another example, to measure whether participants’
Chinese language competency was valued, the original item was “I am well regarded because
of my Chinese language competency”. Further consideration revealed that it actually
contained two questions, namely “are you well regarded or not?” and “what is the level of
your Chinese language competency?” People may be well regarded due to their Chinese
language proficiency or something else, and people may not be well regarded due to their
Chinese language proficiency or something else. It was hard to tell whether people’s Chinese
language proficiency was valued. The original item was then simply reworded as “my
Chinese competency is well regarded”.
Fourthly, enough variance in responses is important for most commonly used statistical
analyses. A split strategy was used to ‘push’ participants to choose different answers in order
126
to get enough variance. Adverbs such as ‘very’, ‘really’, ‘completely’, ‘easily’, and ‘always’
as well as auxiliary verbs such as ‘must’, ‘have to’, and ‘do’ were useful in this regard. By
using these words, participants could be split between those who really agreed and those who
agreed but not that much.
Fifthly, negatives in a statement may lead to errors in responses. It becomes hard for
participants to know whether agreeing with a negatively worded statement might actually
mean they are disagreeing with it. Therefore, negatively worded questions should be restated
or reworded to eliminate this complexity. For example, the original item “I won’t have a good
time if I don’t celebrate Chinese festivals” was first replaced by “I will have a good time if I
celebrate Chinese festivals”. However, the entire sample might be happy to celebrate Chinese
festivals due to the popularity of Chinese culture in Australian society. Therefore, the
participants’ responses may not offer enough variance. Eventually, the item was reworded as
“I feel very sad if I miss celebrating Chinese festivals”.
Finally, jargon needs to be eliminated and replaced by words familiar to all participants.
It is also of great value to work with colleagues and test alternative readings of draft
questions. It is beneficial to check the self-designed questionnaire with individuals
knowledgeable in the topic before using it in the survey. Numerous meetings and discussions
with the supervisory team and experts were held. They have either theoretical knowledge of
the concepts to be operationalised or experience in instrument design and development.
Through meetings and discussions with them, the quality of the instrument improved.
4.3.4 Pilot study
A good research strategy requires careful planning and a pilot study will often be a part of
this strategy. A ‘pilot study’ is a small-scale preliminary study conducted before the main
127
research in order to check the feasibility or to improve the research design (Haralambos &
Holborn, 2004). It is also a trial of a particular research instrument (T. L. Baker, 1994) to
pre-test whether the proposed instrument is appropriate. A pilot study can reveal deficiencies
in the initial design. These can then be addressed before time and resources are expended on a
main study. A pilot study was designed to test logistics and operationalisation in order to
improve the quality and efficiency of the main study. Part of the questionnaire was adapted
from the existing instrument in the literature and part of the questionnaire was designed and
developed by the researcher. It was considered critical to evaluate the feasibility of the
instrument before it could be applied in the main study.
The benefits of the pilot study and its evaluation are two-fold. First of all, face validity
and the appropriateness of the language used for the questionnaire were assessed, according
to the feedback from participants in the pilot study. Ambiguous wording or phrasing was
clarified. Any distress or uncertainty inadvertently elicited by sensitive questions was
decreased to a minimal level by using more comfortable language. Redundant questions were
either deleted or integrated into questions with more succinct language. Inadequacies in the
instructions were also identified. Secondly, reliability testing included in the pilot study is a
prerequisite for reliable findings. Cronbach’s alpha estimates the internal consistency of a set
of items and can be used as a measure of reliability. Though the alpha value is item number
and sample size dependent, an alpha level higher than .80 indicates that items are reliable and
the set of questions is internally consistent (Kline, 1999). An alpha level lower than .80
indicates that there may be problems with internal consistency. This procedure represents a
refinement by identifying problematic items in the test. Though deleting these items from the
questionnaire may affect the validity, it will increase the internal consistency of the
instrument. Apart from the alpha value, corrected item total correlation is another relevant
measure. Its value lower than .33 indicates that less than 10% of the variance in the scale is
128
accounted for by that corresponding item (R. Ho, 2006) and that this item is of concern. It is
also useful to determine whether Cronbach’s Alpha value increases if a particular item is
deleted. In short, the pilot study serves to plumb the validity and reliability of the instrument.
A pilot study is normally small in comparison with the main study. It requires a smaller
number of participants in contrast to the main study. The sample size for Cronbach’s
coefficient alpha test could be as small as 30 (Duhachek, Coughlan, & Iacobucci, 2005;
Yurdugul, 2008). Convenience sampling is widely used to get an inexpensive approximation
of the truth (Walonick, 1993). As the name implies, the sample is selected because it is
convenient for the researcher to approach participants. The participants are often known to
the researcher and are willing and available to be studied. This non-probability method is
often used in research to get a gross estimate of results, without incurring the cost or effort
required to select a random sample. Therefore, 38 participants were approached by
convenience sampling for the pilot study. The participants in the pilot study were excluded
from the final sample of the main study, because their participation in the pilot study could
have influenced their behaviour in the main study (Haralambos & Holborn, 2004).
4.3.5 Main study
Sample size
Different statistical models have different rules to identity the optimal sample size required to
produce reliable results. SEM is the major statistical model applied to the data analysis in the
quantitative phase. SEM in general requires a large sample size, which varies as a function of
choice and value of fit index, number of variables / degrees of freedom, relation among
variables, and statistical power (Kim, 2005). Consulting the algorithm developed for sample
size planning in SEM (Kelley & Lai, 2011; Kim, 2005) and the conventional minimal sample
129
size required for SEM (Bentler & Chou, 1987), this study was conducted with the intent of
gaining 200 complete cases as the sample size for SEM.
Multiple regression models were also constructed in this study. A rule-of-thumb that
provides a rationale to estimate sample size for regression analysis is proposed by Green
(1991). When significance level and power value were set at .05 and .80 as appropriate for a
wide range of behavioural research (Cohen, 1992; Hair et al., 2006), the functional
relationship between sample size (N) and effect size (f2 or R
2) as well as the number of
predictors (k) was established (Green, 1991). Although Cohen (1988) argues that the choice
of values for effect size (f2 or R
2) should depend on the research area, he proposes, as a
convention, f2 values of .02, .15, and .35 to serve as operational values for small, medium,
and large effect size respectively, when the corresponding R2 values are .02, .13, and .26
(Cohen, 1988). The functional equation is shown below:
N ≥ L / f2, where f
2 = R
2 / (1 - R
2) and L = 6.4 + 1. 65 × k - .05 × k
2
In the regression analysis, there were two predictors, namely ‘capital’ and ‘Chineseness’.
Accordingly, L valued at 9.5 (L = 6.4 + 1. 65 × 2 - .05 × 22 = 9.5). For a medium effect size
of .15, the minimum sample size for the regression analysis was 63 (N ≥ 9.5 / .15 = 63). This
value was verified by using the software G-Power. When significance level, power value, and
effect size were set at .05, .80, and .15, G-Power reported the calculated sample size of 68 as
the minimum sample size required by the regression analysis, which was very close to the
manually calculated value 63. Taking account of the minimal sample sizes required by SEM
and multiple regression, the study sought at least 200 complete responses in order to produce
reliable results.
130
Sampling
It is characteristic of most sampling techniques in the quantitative approach that the aim is to
provide information about the total population. The point of sampling techniques is that
conclusions regarding the total population and not just the sample can be drawn with a
recognised degree of uncertainty (Danermark et al., 2002). Therefore, it is imperative to
sample in a careful and explicit manner.
Every sampling method has its strengths and weaknesses. The selection of sampling
method is contingent upon the desired sample size, the characteristics of the target population,
and the availability of participants (Creswell, 2008). Simple random sampling and systematic
sampling could not be used in the current study because it was impossible to identify each
member in the target population. Stratified sampling and cluster sampling are often used
when there is a target population that is large and easy to access. Though Chinese constitute
the third largest group among all immigrants in Australia, just behind people from the United
Kingdom and New Zealand, Chinese Australians only account for 4.3% of the Australian
national population (ABS 2011 Census) and they are a numerical minority. This increases the
difficulty of sampling because the smaller the population, the harder it becomes to approach
the participants to represent that population. As Groves et al. (2004) argues, what makes a
population rare sometimes is not its absolute size but its size relative to other coexisting
populations. In this regard, stratified sampling and cluster sampling were not possible.
Snowball sampling avoids the challenges that applying other sampling methods raises
in the real world. It occurs when the researcher asks participants to recommend other
individuals to participate in the study. It is commonly used when the participants are rare and
thus difficult to access or identify (Nardi, 2006), for example, in heritage research (Gibbs &
Hines, 1992; Hall, 1992; Kiang, 2008; Pao, Wong, & Teuben-Rowe, 1997; Root, 1992).
Therefore, snowball sampling was used to approach participants in the current study.
131
To ‘get the snowball rolling’, a request was posed in the participation statement of the
online questionnaire and through informal conversation with individuals through personal
networks, asking participants to invite other participants whom they know and meet the
participation requirements. This led to sampling of individuals who had not been anticipated
when the project began because the snowball sampling proceeded after the study began.
Although the research started with a relatively short list of participants, the list grew like a
snowball as names were added through the referral of participants. Eventually, snowball
sampling netted 230 respondents. A detailed description of these participants is presented in
Chapter Five.
Data analysis
The instrument for the main study was the refined and modified version of the original online
questionnaire trialled in the pilot study. Responses to the online questionnaire provided by the
participants were transferred directly into SPSS to produce the raw data set. Given this
electronic process of data production, data entry error should not be a concern. In the raw
data set, each row recorded all the variables corresponding to each participant. Each line
recorded each variable across all the participants. The production of single-factor congeneric
measurement models and structural models investigating the relationships amongst constructs
is pivotal to successful analysis. To this end, SEM was the major statistical technique used in
this study. IBM SPSS AMOS 21 was used to run SEM. The advantages of SEM are reported
below.
First and foremost, SEM is theory-driven. Bourdieu’s theory of habitus, capital, and
field (Bourdieu, 1989a) argues that practice results from the relations between one’s
dispositions (habitus) and one’s position in a social world (capital), within the current state of
132
play of that social arena (field). Specifically, Bourdieu (1991) theorises that people make
choices about language according to the amount of different resources (capital) as well as the
dispositions (habitus) that they have within given social fields. As argued in Chapter Three,
this study extended Bourdieu’s explication of agents’ language choices within a particular
language to agents’ choices from different languages. SEM was applied to test the complex
entanglement amongst the habitus of ‘Chineseness’ of young Chinese Australian adults,
various forms of capital available to them, and their CHL proficiency resulting from their
CHL practice. SEM provides the justification for the specification of this theoretical
dependence relationship. The theory-based approach to SEM is a distinct strength of this
technique, as it entails a mode of thinking that specifies the theoretical framework more
exactly, tests the theory more precisely, and yields a more thorough understanding of the data
(R. Ho, 2006).
Secondly, SEM has the capability to assess how well the indicators reflect their
respective latent variables by Confirmatory Factor Analysis (CFA) and how well the latent
variables are related in a series of dependence relationships between ‘exogenous variables’
and ‘endogenous variables’. An exogenous variable is one whose variability is assumed to be
determined by factors outside the model under consideration; an endogenous variable, on the
other hand, is one whose variability is to be explained by exogenous and other endogenous
variables within the current model (Hair et al., 2006). SEM can assess how well each
theoretical concept is operationalised by their corresponding questionnaire items and how
well these theoretical concepts are related as respective exogenous and endogenous variables.
Moreover, SEM improves statistical estimation by accounting for unique disturbances
in the estimation process (R. Ho, 2006). Many other statistical models assume that variables
in the analyses are error-free. However, concepts can seldom be measured perfectly, from
either theoretical or practical perspectives. This is either ascribed to external errors associated
133
with inaccurate responses by the participants, or attributed to internal errors associated with
imperfect operationalisation of the concepts. Consequently, measured variables usually
contain at least moderate amounts of error (R. Ho, 2006). However, in SEM, scores on the
indicators are used to develop estimates of their respective latent variables. As these estimates
are derived on the basis of the common or shared variance among the indicators, scores on
the latent variables are unaffected by measurement errors (R. Ho, 2006). Therefore,
measurement errors are removed and offset in SEM.
It is important to distinguish a reflective model from a formative model before an SEM
is applied because SEM applied to formative models can lead to large standard errors and
unstable parameter estimates. According to Coltman, Devinney, Midgley, and Venaik (2008),
reflective models are distinct from formative models in the following ways. In reflective
models, causality runs from the latent variable to its indicators. Causality runs in an opposite
direction in formative models. As a result of causality direction, variation of the latent
variable causes the variation of the indicators in reflective models, while it is the other way
around in formative models. In reflective models, the latent variable exists as a theoretical
construct independent of the indicators used to measure it. Indicators are represented by the
latent variable. In formative models, the latent variable is formed as a combination of its
indicators. Indicators define the latent variable. Consequently, indicators in reflective models
are interchangeable. Adding or dropping an indicator does not change the conceptual domain
of the theoretical construct. This is not the case in formative models. Moreover, indicators in
reflective models share a common theme and they should have high positive intercorrelations.
In contrast, this is not necessarily the case in formative models. Therefore, these two models
are conceptually, substantively, and psychometrically different. Bearing this in mind, the
design of the online questionnaire has taken careful consideration of the basis for the
construction of SEM. Questionnaire items did not function to form and define their
134
corresponding theoretical constructs. Rather, they were designed to reflect the attributes
associated with their corresponding theoretical constructs. As a result, variance of the existing
constructs caused that of the item measures. In line with the above perspectives, reflective
models were established and SEM applied to these reflective models was considered
appropriate for the current study.
Two critically important assumptions associated with SEM are: data are measured at an
interval level and have a multivariate normal distribution (Byrne, 2001). It has been argued
earlier that the data measured by a 7-point-Likert-type scale are widely accepted as proxy
interval level of measurement in practice, which satisfies the first assumption. However, most
data in practice fail to meet the assumption of normal distribution and the presence of
excessive kurtosis is particularly troublesome for the normality assumption (Byrne, 2001). As
a rule of thumb, if the value of skewness (kurtosis) falls beyond the range from minus twice
standard error of skewness (kurtosis) to plus twice standard error of skewness (kurtosis), the
univariate normal distribution assumption is considered to be violated (Field, 2009). That is,
to satisfy the univariate normal distribution assumption, the critical ratio between the value of
skewness (kurtosis) and that of the standard error of skewness (kurtosis) should be less than
|2|. Alternatively, a Kolmogorov- Smirnov Test can be used to test the normality. A significant
Kolmogorov-Smirnov value indicates the violation of the normality assumption. Even if all
univariate distributions are normal, the joint distributions of the variables may depart
substantially from multivariate normality (Stevens, 2002). Mardia’s coefficient was used as
an indicator of the degree of multivariate normality. A Mardia’s coefficient greater than 3
implies the violation of multivariate normality (Yuan, Marshall, & Bentler, 2002). One
approach to handling the presence of non-normal data is to use a procedure known as
‘bootstrapping’ (Byrne, 2001). The decision as to whether or not bootstrapping needs to be
used will be based on the degree of multivariate normality indicated by Mardia’s coefficient.
135
SEM is a multivariate technique that can be described as a combination of CFA and
path analysis. Correspondingly, SEM consists of two parts: the ‘measurement model’ and the
‘structural model’. The measurement model is concerned with the relations between
indicators and latent variables (R. Ho, 2006). It represents the degree to which the indicators
capture the essence of the latent variable (Hair et al., 2006). It specifies the relations between
a set of questionnaire items and the theoretical concepts these items were designed to
measure. This is achieved by CFA. The structural model is a flexible and comprehensive
model that specifies the pattern of relationships among latent variables, namely exogenous
and endogenous variables (R. Ho, 2006). This will be achieved by path analysis. Both models
were distinctive in the current study. The measurement model provided a test for validity and
reliability of the questionnaire items employed to measure the theoretical concepts. This
further improved the validity and reliability of the instrument after the pilot study. The
structural model offered a direct test of the theoretical framework and hypotheses, and led to
the answers to RQ1. The rationale for the initial measurement model and the subsequent
structural model are explicated below.
The proper specification of the measurement model is necessary before meaning can be
assigned to the analysis of the structural model. Therefore, a measurement model was firstly
specified. The indicator variables or the questionnaire items were assigned to the
corresponding latent variables or the theoretical constructs. The overall model fit of the
measurement model was assessed by CFA. Several approaches were used to check the overall
model fit. First of all, the chi-square goodness-of-fit statistic was computed. The chi-square
test assesses an exact null hypothesis that the data included in the model do not significantly
deviate from the population. A chi-square value corresponding to an insignificant p value
indicates that there is not enough statistical evidence to reject the null hypothesis, and
therefore demonstrates an acceptable model fit. However, chi-square value is sensitive to
136
sample size and increases as a direct function of sample size. In cases of large samples
needed by SEM, the power of the statistical test underlying the SEM approach is very high.
With a large sample size and a great deal of statistical power, almost every reasonable model
will be rejected if only the chi-square value and the associated p value are considered (R. Ho,
2006). Bentler (1990) advises against the sole use of the chi-square value in judging the
overall fit of the measurement model. To this end, absolute fit measures, such as RMSEA
(Root Mean Square Error of Approximation), and some incremental fit measures, such as CFI
(Comparative Fit Index), IFI (Incremental Fit Index), NFI (Normed Fit Index), RFI (Relative
Fit Index), and TLI (Tucker-Lewis Index) are also computed to augment the chi-square
goodness-of-fit. Among these, CFI and TLI are reportedly used most often (Hair et al., 2006).
Although there are no clearly established rules as to what constitutes a good fit, a widely
applied guideline suggests that the incremental fit measures should be higher than .90 to
indicate a good overall fit of measurement model (Bentler, 1990; Hair et al., 2006) and that
RMSEA values below 0.08 are deemed acceptable to represent the goodness-of-fit when the
proposed model is estimated in the population (Hair et al., 2006; R. Ho, 2006).
CFA also provides additional diagnostic information that may suggest modifications for
improving the model. These diagnostic cues include regression weights, standardised
residuals, and modification indices. To test whether each item in the questionnaire is
significantly represented by their respective latent variable, the unstandardised and the
standardised regression weights are calculated. The unstandardised regression weights are
checked by the critical ratio and its corresponding significance level. A significant (p < .05)
critical ratio indicates the corresponding item in the questionnaire is a significant indicator of
the respective latent variable, and vice versa. The standardised regression weights should be
at least .5 and preferably .7 to consider the corresponding item as a significant indicator of the
respective latent variable (Hair et al., 2006). Typically, standardised residuals less than |2.5|
137
do not suggest any problem (Hair et al., 2006). Conversely, residuals greater than |4.0| serve
as a warning and suggest a potentially unacceptable problem (Hair et al., 2006). Standardised
residuals between |2.5| and |4.0| deserve some attention but may not suggest any changes to
the model if no other problems are associated with these two items (Hair et al., 2006).
Modification indices demonstrate how much the chi-square value would be reduced.
Modification indices of greater than |4.0| suggest that the model fit would be improved by
allowing corresponding variables and error terms to correlate. However, given the strong
theoretical basis of SEM, changes to the measurement model cannot be solely based on the
diagnostics provided by CFA. The decision whether or not to implement these modifications
rests on the theoretical justification (R. Ho, 2006) along with those suggested diagnostics
(Hair et al., 2006). As mentioned earlier, SEM is a theory-based approach. Without strong
theoretical justification, employing the modification to improve model fit only capitalises on
the uniqueness of the data set but the results will most likely be atheoretical.
Traditional approaches to reporting reliability and validity are not easily transferred to
SEM because they do not take account of the congeneric nature of the model and assume that
the factor loadings are considered equal. As a result, traditional approaches underestimate
both reliability and validity measures (Graham, 2006). To offset the drawbacks of traditional
approaches, four measures of reliability, namely Squared Multiple Correlations (SMCs),
construct reliability, variance extracted, and coefficient H, as well as two measures of validity,
namely construct validity and convergent validity were reported. The rationale behind these
reliability and validity measures will be explained in Chapter Five.
After checking and improving the fit of the measurement model as well as confirming
the reliability and validity of the measurement model, the structural model that assesses the
relationships amongst latent variables can be constructed. The regression weights matrices
demonstrate how strong and significant the relationships amongst these latent variables are.
138
In this study, the structural model talked back to the theoretical framework, tested the two
hypotheses, and helped to answer RQ1. Attuned to the theoretical framework, the following
two structural models (see Figures 4.2 and 4.3) were constructed for hypothesis testing.
Straight arrows depict a dependence relationship, which means the impact of one variable on
another. In a measurement sense, dependence relationships exist from latent variables, or
theoretical constructs, to their indicators or the questionnaire items by which latent variables
are measured. In a structural sense, dependence relationships occur between latent variables,
with arrows flowing from the exogenous variable to the endogenous variable. In Figures 4.2
and 4.3, latent variables are represented by ovals while indicators are represented by squares.
‘I’ stands for ‘indicator’ or ‘item’ in the questionnaire and ‘e’ stands for measurement error or
more correctly ‘unique disturbance’.
Figure 4.2. Structural relationship between ‘Chineseness’ and ‘CHL proficiency’
Figure 4.3. Structural relationship between ‘capital’ and ‘CHL proficiency’
I1 I2 I3 …
e e e
Chineseness
I1 I2 I3 …
CHL proficiency
e e e
e
e
Cultural
capital
e
Capital
I1 I2 I3 …
e e e
CHL proficiency
Symbolic
capital
Social
capital
Economic
capital
e e e
139
4.4 Qualitative phase
The initial quantitative phase of this project sought general explanations and systematic
relations within the phenomena under investigation. Accordingly, the quantitative phase was
not and could not be concerned with the individual and exceptional case, as well as
participants’ detailed and deep perceptions and viewpoints. To tackle these dimensions of the
phenomena, the subsequent qualitative phase was designed to function as a complement of
the initial quantitative phase.
Qualitative research seeks to gain a deep holistic perspective and understanding of
people’s experiences by asking broad and general questions (Fraenkel & Wallen, 1996). The
basic principle of a qualitative study is that there are in-depth meanings embedded in people’s
experiences and that these meanings are mediated through researchers’ own perceptions
(Merriam, 1998). As such, qualitative research is concerned with the nature, explanation, and
understanding associated with people’s experiences, views, and beliefs. Based on the findings
of the quantitative analysis, more analysis working with qualitative data is needed to extend,
probe, and illustrate the results of the quantitative phase. The initial quantitative phase
examines the research problems of whether the habitus of ‘Chineseness’ of young Chinese
Australian adults and their investment of different forms of capital can produce their CHL
proficiency. Subsequently, the qualitative phase addresses RQ2: How do young Chinese
Australian adults understand their CHL learning in relation to (potential) profits produced by
this linguistic capital in given fields?
In order to answer this question, the qualitative phase adopted an interview approach. A
research interview is a purposive conversation (Lodico, Spaulding, & Voegtle, 2006) and an
interchange of views between the interviewer and the interviewee(s) on a topic of mutual
interest, which constructs sites for knowledge production (Kvale, 1996). It is the most
common data collection method in qualitative research, particularly for studies like the
140
current project, when intimate, repeated, and prolonged involvement in the life and
community of the participants is not possible (Lodico et al., 2006). Thus, interviews become
the most pragmatic and productive way to understand what particular experiences mean to
participants.
Multiple interviews allow participants to reflect on their experiences and to add to what
they said in earlier interviews (Lodico et al., 2006). In addition, new questions may arise after
the first interview and a second interview is required (Bourdieu, 1999c). In order to explore
the deep meaning embedded in the participants’ experiences, multiple interviews were used to
collect the qualitative data for this study.
Interviews have often been criticised for lack of objectivity, due in particular to the
human interaction inherent in the interview process. Kvale (1996) argues for three
conceptions of objectivity suitable for interview research: freedom from bias, intersubjective
knowledge, and reflecting the nature of the object. Firstly, objectivity as freedom from bias
refers to “good, solid, craftsmanlike research” (Kvale, 1996, p. 64), producing reliable
knowledge that has been systematically cross-checked, controlled, and verified, undistorted
by personal bias and prejudice as far as possible. Secondly, this knowledge must be
intersubjectively testable and reproducible, which means that repeated research of the same
problem by different researchers should yield the same data. This is what Kvale (1996, p. 64)
refers to as “dialogical intersubjectivity”, which is an agreement through a rational discourse,
reciprocal critique, and communicative validation among researchers, as well as between
researchers and their participants. Thirdly, objectivity also means the reflection and
expression of the nature of the participants researched by treating the participants as the
speaking subjects existing in a linguistically constituted and interpersonally negotiated
interview process. As such, the interview, as a linguistic, interpersonal, and interpretive
method, becomes more objective.
141
To maximise these dimensions of objectivity, the interview also requires attention to
two methodological concerns, namely ‘reliability’ and ‘validity’. ‘Reliability’ refers to how
consistent the results are, and ‘validity’ means whether an interview investigates what is
intended to be investigated (Kvale, 1996). These two issues do not belong to some separate
stages of an investigation, but should be addressed throughout the entire interview research
process, starting from thematising, designing, interviewing, transcribing, and analysing, to
reporting. Each stage will be discussed in the following sections.
4.4.1 Thematising
Thematising refers to a conceptual clarification and a theoretical analysis of the theme(s) to
be investigated, and informs the formulation of the questions to be asked (Kvale, 1996). The
theoretical framework developed in Chapter Three (see Figure 3.4) has conceptualised the
main themes under investigation and has related these themes, or theoretical constructs, to
one another. As Bourdieu (1991) suggested, languages can produce profit in response to the
demands of a given field. Accordingly, the overarching theme to be investigated in the
qualitative phase is in relation to the profits produced through young Chinese Australian
adults’ CHL learning in certain fields. Attuned to this theme, the analytical questions driving
the qualitative phase include: Does young Chinese Australian adults’ investment in CHL
learning have any return at a given time and place? Does CHL proficiency produce any
valuable asset for young Chinese Australian adults in any situation? Does lack of CHL
proficiency impact on their life in any way? These questions align with the research interview
approach. They can return to the subjective experience of young Chinese Australian adults
regarding their CHL usage and learning trajectories over time. The research interview can
obtain comprehensive descriptions about the CHL usage and learning experience from the
142
participants’ perspective, and then provide the basis for a deep analysis to portray the essence
of the experience, the essential meanings of CHL to these participants.
If the structural models (see Figures 4.2 and 4.3) deduced from the first part of the
theoretical framework (see Figure 4.1) functioned to drive the initial quantitative phase, the
subsequent qualitative phase was guided by the second part of the theoretical framework (see
Figure 4.4) deduced from the overall theoretical framework (see Figure 3.4). It is worth
noting that the single-headed arrows demonstrated in Figures 4.1 and 4.4 were in the opposite
direction because the research questions in the initial quantitative phase and the subsequent
qualitative phase addressed relationships that went in different directions.
Figure 4.4. Theoretical framework guiding the qualitative phase
4.4.2 Designing
Designing consists of overall planning and preparing the methodological procedures for
obtaining the intended knowledge (Kvale, 1996). The following sections explain the rationale
behind the design for interview and sampling strategy in the qualitative phase.
CHL proficiency as
linguistic capital
Four forms of
capital
Chineseness as
habitus
Fields
143
Interview design
The one-to-one interview is purported to be the most commonly used data collection method
in qualitative social research (Ryan, Coughlan, & Cronin, 2009). It is a data production
process of asking questions and recording answers from only one participant at a time. It is a
valuable approach to gain insight into participants’ perceptions, understandings, and
experiences of a given phenomenon under certain circumstances (Lodico et al., 2006; Ryan et
al., 2009). The qualitative phase aimed to address participants’ perspectives and viewpoints
on their own CHL commitments, and thus a one-to-one interview method was selected.
Semi-structured interviewing provides the flexibility to combine an interview schedule
of prepared questions on predetermined topics with explorations of any unanticipated
responses or issues that emerge through the use of open-ended questioning (Tod, 2006). It has
a sequence of themes to be covered, as well as suggested questions, yet at the same time, it is
open to changes of sequence and further probes in order to follow up on answers given and
experiences explained by the participants. The flexibility of the semi-structured interview
allows the interviewer, who is the researcher himself in the current study, to pursue a series of
less structured questions and to explore unforeseen issues raised spontaneously by the
participants. Reponses to structured questions can net useful information. The open-ended
responses can explore details that may surface during the course of the interview. This
facilitates the production of richer and more textured data from the participants than from
formally structured questions in standardised interviews. Therefore, a semi-structured
interview design was adopted in the qualitative phase.
If residing in Brisbane, participants were interviewed face-to-face. Otherwise, they
were interviewed by telephone or via online software such as Skype or MSN, because
telephone and online interviews were more cost-and time-effective than face-to-face
interviews in this situation. All the participants have achieved native, native-like, or nearly
144
native-like command of English. In contrast, their Chinese functioned as their HL, the
proficiency of which was lower or much lower than that of their English. Therefore, the
interviews were conducted in English.
An interview protocol was prepared for asking questions and recording key points
during the interview. As suggested by Lodico, Spaulding, and Voegtle (2006), the interview
protocol should usually include a heading that records date, place, and interviewee code, a
brief introduction of the study, the preliminary questions to be asked, and possible probing
questions to follow up, or in Bourdieu’s term “constant improvisation of pertinent questions”
(Bourdieu, 1999c, p. 613). These interview questions have essential relationships with the
research questions. However, interview questions tend to be more contextual than research
questions. Their development requires creativity, sensitivity, and insight, rather than a
mechanical translation of the research questions into an interview guide.
Patton (1990) offers advice concerning the development of good interview questions to
address reliability and validity. Questions can be asked from a variety of angles, including
experience questions, opinion, feeling, or value questions, knowledge questions, and
background or demographic questions. Another piece of advice is to offer a quotation
selected from another source that contains ideas on which the participants may comment. The
value of the quotation is that it attributes to someone else ideas that are usefully provocative
but are not in the researcher’s own voice. In contrast to the above good practice for interview
questions, questions that are dichotomous ‘yes or no’ questions are not helpful for exploring
details because such questions precipitate short answers. Multiple questions should be
avoided because the participants will most likely talk more fully about one of the several
questions and forget the others.
Leading questions that suggest an answer or contain information expected by the
researcher (Patton, 1990) have pros and cons. On the one hand, the leading effects of such
145
questions are well documented (Kvale, 1996). They give the participants hints about what
would be considered a desirable or appropriate kind of answer. As a result, the answer may
not reflect the thoughts of the participants. On the other hand, they can be judiciously used to
clarify the exact meaning of the participants. The decisive issue for interview questions is not
whether to lead or not, but where the questions do lead, whether they lead in important
directions that yield new and worthwhile knowledge (Kvale, 1996).
Structured within the themes and guided by the research questions, the interview
questions in the current qualitative phase consisted of four sections, namely CHL learning,
CHL usage, attitudes towards CHL, as well as particular attributes for each participant (see
Appendix 4). The prepared probing questions were developed with reference to the
theoretical framework of the current study. Bourdieu’s concept of capital, habitus, and field
informed the probing questions. These probing questions were used to elicit information to
explore whether CHL proficiency can produce ‘capital’ and ‘habitus’ for the interview
participants in given fields, however these theoretical terms were expressed in everyday
terms, such as ‘resources’, ‘Chinese identity’, and different ‘contexts’ or ‘situations’.
To make sure that the interview questions made sense to the participants and the
participants felt comfortable with the interview questions, a pilot interview was conducted
before the main interview. The pilot interview served to reshape the original interview
questions and interview style. For example, in the original design one of the questions was “If
you were to have children, would you encourage them to learn Chinese?” One of the pilot
interviewees suggested that to certain extent, this was a leading question because anyone with
a Chinese ancestry tends to say yes to this question even if Chinese language is not the first
choice for their children. Following this suggestion, the original question was amended to “If
you were to have children, would you encourage them to learn another language other than
English?” If the participants said “Chinese”, the reasons behind this choice were probed. If
146
the participants said any other language, they were asked why they would not encourage their
children to learn Chinese. In this way, the primary and overall question was presented at the
very beginning and then the interview process was managed by using interview skills to elicit
information. When the information elicited appeared to be drying up, pre-prepared probe
questions planned on the interview protocol were used. As Cavana, Delahaye, and Sekaran
(2001) suggested, a semi-structured interview should commence with an open-ended question
then progress to a more closed-ended form. The pilot interview provided good practice to
improve the researcher’s interview skills. It also prompted the researcher to rework some of
the questions into a more open form.
Audio-recording was used to record the interviews. Two audio recorders were used in
case of equipment failure. In addition to the audio recording, field-notes were only used when
necessary to record key points of participants’ responses as faithfully as possible and to serve
as reminders for the researcher to return to any issue that needed more probing in the
interview. Therefore, digital recording and field-notes were used together to ensure high
quality data.
Before the interview, appropriate locations and time were found. A time block
convenient for the participants and available for both the researcher and the participants, as
well as a quiet and physically comfortable location were the overriding considerations.
Participants’ willingness to cooperate can be contingent on convenience. Therefore, it was
important to defer to the participants’ needs as much as possible.
At the very beginning of each interview, the context was introduced by a briefing about
the purpose of the interview, ethical considerations, the use of audio recorders, and the
general topics in the interview. Each participant was given an information sheet to read prior
to the interview (see Appendix 2). If participants decided to take part in the study, the
researcher asked them to sign a consent form (see Appendix 3). In this way, the researcher
147
introduced the proposed research to the participants and obtained their informed consent in
compliance with ethical requirements. Before starting the interview, the researcher also asked
whether the participant had any questions. Such briefing in the first minutes of an interview is
crucial (Kvale, 1996). This was a good opportunity to establish a friendly atmosphere and
mutual trust by attentive interactions, where the participants could have a sense of who the
researcher was before they talked freely about their lived world to a stranger, and where the
researcher could show interest, understanding, and respect for what the participants said.
By the end of the interview, the participants have given much information about their
lived world. They might have obtained new insights into aspects of their lived world but
sometimes they might not have received anything in return. Therefore, the interaction was
rounded off by the researcher reflecting on some of the main points learned from the
participants. The participants might want to comment on this feedback. The interaction was
thereafter concluded by the researcher saying, for example, “I have no further question. Do
you have anything more you want to bring up, or ask about, before we finish the interview?”
(Kvale, 1996, p. 128). This gave the participants an additional opportunity to deal with issues
they have been thinking or worrying about during the interview. It was worthwhile for the
researcher to set aside a short block of quiet time right after each interview to recall and
reflect on what has been learned from the particular interview. The immediate impressions, in
the form of field notes or audio records, provided a valuable resource for later analysis of the
transcripts.
Sampling
There is no definitive answer to the question of how many participants are needed in an
interview study (Mertens, 2005; Mertler & Charles, 2005). Within qualitative research,
148
although diverse samples might provide a broader range to distil the essence of the
phenomenon, data from only a few participants who have experienced the phenomenon and
can provide a detailed account of their experience about the phenomenon might suffice to
illuminate its core elements.
Unlike the sampling principles in the quantitative approach, where the samples usually
are statistically representative, the sampling principle in the qualitative approach is more
strategic. Therefore, purposeful sampling is more often used in qualitative research (Lodico et
al., 2006). The logic and power of purposeful sampling is the selection of cases that provide
rich and deep information central to the purpose of the research (Lodico et al., 2006; Patton,
1990). Participants in the initial survey were also invited to participate in an interview. Those
who were interested in the interview and who left their contact details became candidate
interview participants. The current study adopted purposeful sampling to select participants
from the interview candidate list.
Specifically, Patton’s (1990) extreme or deviant sampling and maximum-variation
sampling was used. Under the extreme or deviant sampling strategy, interviewees can be
selected from extreme cases that are unusual or special in some way. The designations imply
that they are odd phenomena or cases that represent the extremes. These types of cases often
provide considerably more insight than representative or average cases (Danermark et al.,
2002). In the qualitative phase, extreme cases were identified by revisiting the outliers in the
initial quantitative phase. Identification strategies of the outliers will be explained in Chapter
Six. Outliers who were also candidate participants were invited to participate in the interview.
New knowledge learnt from an intensive examination of the experiences of special or unusual
participants through the interview strategy in the qualitative phase were considered to offer a
good complement to the knowledge learnt from statistical depictions of what the more regular
cases were like through the survey in the initial quantitative phase.
149
Interviewees were also selected on the criterion of maximising variation. They were
selected from varied cases to gain insight into the importance of various conditions for
producing the particular phenomenon under investigation. This strategy can document what is
unique about each situation as well as what is common across diverse settings. In the
qualitative phase, participants’ self-reported CHL proficiency in the initial quantitative phase
was taken into account. To cover maximum variation, candidate participants with highly
divergent levels of CHL proficiency were identified and invited to participate in the interview.
Account was also made to maximum variation in participants’ demographic features.
Eventually, Patton’s (1990) extreme or deviant sampling and maximum-variation sampling
helped to net five interviewees. A detailed description of these interviewees is presented in
Chapter Six.
4.4.3 Interviewing
Interviewing is an engaging process, an interpersonal situation, and an interactional
conversation (Kvale, 1996). It is up to the researcher to create in a short time an interpersonal
situation that allows the interaction to go beyond merely a polite conversation or exchange of
ideas. The researcher must establish an atmosphere where the participants feel safe and
comfortable enough to talk freely about their lived world. In order to achieve this, several
important issues need due attention. They are addressed as follows.
There is a definite asymmetry of power during the interview process because it is the
researcher who defines the situation, introduces the topics of the conversation, and steers the
course of the interview through further questions. Bourdieu (1999c) has explained this
asymmetry:
150
It is the investigator who starts the game and sets up its rules, and is usually the one
who, unilaterally and without any preliminary negotiations, assigns the interview its
objectives and uses…This asymmetry is reinforced by a social asymmetry each time
the investigator occupies a higher place in the social hierarchy of different types of
capital, cultural capital in particular. (p. 609)
Due to the unequally distributed cultural capital, there was an asymmetry of power existing in
the context of the current study, where the researcher, an educated native Chinese speaker,
had conversations around topics associated with CHL learning experiences of the Chinese
Australian participants. In this social field, the participants might assume that the researcher
had better Chinese language proficiency or more cultural capital in Bourdieu’s sense than
themselves. This assumption entailed hierarchical field positions occupied by the researcher
and the participants based on different levels of Chinese language competence valued as
cultural capital in this field. Because of the recognised value, Chinese language proficiency
might accrue symbolic capital. Lack of this symbolic capital might yield embarrassment of
‘loss of face’ and generate an uncomfortable environment for the participants during the
interviews.
In order to control the effects of asymmetry of power, the “symbolic violence” exerted
through the asymmetric relationship should be reduced as much as possible (Bourdieu, 1999c,
p. 609). In practice, naivety characterises the researcher’s special learner role (Kvale, 1996).
It entails a frame of mind to set aside any assumptions that the meaning of the participants
has been known to the research. Such assumptions would preclude the researcher from
seeking explanations and shut down further probes. The learner role can be administered
through verbal “signs of feedback” or “response tokens” (Bourdieu, 1999c, p. 610), such as
“yes”, “right”, “oh”, and “ok”, as well as aiding explanations in a proposed rather than an
151
imposed way, such as suggestions to offer multiple and open-ended continuations to the
participants’ hesitations or searchings for appropriate expressions (Bourdieu, 1999c, pp.
614-615); and non-verbal means, such as an open posture, approving nods, appropriate facial
expressions, and good eye contact. These verbal and bodily signs of attention, interest,
approval, encouragement, and recognition are conditions for a continuing interchange that
encourages the collaboration of the participants interviewed. These practices ensure that the
researcher adopts the role of a listener and a seeker of knowledge during the whole process of
interviewing. In Bourdieu’s words, it is “active and methodical listening” that leads to
“adopting the interviewees’ language, views, feelings, and thoughts” (Bourdieu, 1999c, p.
609) and signals “the interviewer’s intellectual and emotional participation” (Bourdieu,
1999c, p. 10).
Apart from “active and methodical listening”, “proximity and familiarity” also provide
conditions of “nonviolent communication” during the interview (Bourdieu, 1999c, p. 610).
Common cultural and physical dispositions shared by the researcher and the interviewees
may contribute to this proximity and familiarity. For example, William Labov (1972) asked
young blacks to conduct the linguistic investigation of the speech used by blacks in Harlem.
In the same way, Bourdieu (1999c, p. 611) gave “training in interview techniques to
individuals who could have the kind of familiar access to certain categories of respondent”.
In the current study, the researcher, who is a native Chinese, is culturally and physically close
to the interviewees, or in Bourdieu’s sense “linked to them by close familiarity” (1999c, p.
611). When a young Chinese researcher interviewed these young Chinese Australians, the
conversation could spring from the dispositions attuned to each other. The interviewees
seemed to consider this situation as an exceptional opportunity to “make themselves heard”
and to “carry their experience over from the private to the public sphere” (Bourdieu, 1999c, p.
615). Their speech seemed to convey “a joy of expression” (Bourdieu, 1999c, p. 615).
152
However, this relatively easy access and affinity with the interviewees do not necessarily lead
to biased interpretation of the data to meet the researcher’s own assumptions and expectations.
As an international student from China, the researcher was more confined to the university
life and removed from the more complex lived worlds of Chinese Australians. Such social
difference enabled the researcher to step back as a listener from the interviewees’ experiences.
In brief, ‘proximity’ and ‘social difference’ avoid the two problematic extremes: “total
divergence” between the researcher and the interviewees, where understanding and trust are
not possible, and “total overlap”, where nothing can be said and questioned because
everything goes without saying (Bourdieu, 1999c, p. 612).
To sum up, the interview is a social interaction where the relationship between the
researcher and the interviewee(s) is of paramount importance. The researcher has the
responsibility of ensuring that the interviewee is at ease and also of taking the role of an
active learner. A relaxed, confident, and attentive approach demonstrated by the researcher
helps a good interview relationship to develop. However, the interview is more than just a
social interaction. It is a process with a distinct purpose to derive meaning from a particular
situation. It is a performance that demands skill and forethought. In Bourdieu’s sense,
interviewing demands a “craft”, a real disposition to pursue truth, which disposes the
researcher to improvise on the spot, in the urgency of the interview, strategies of
self-presentation and adaptive responses, encouragement and opportune questions to help
participants deliver up their truth (Bourdieu, 1999c, p. 621). It is “a sort of spiritual exercise
that, through forgetfulness of self, aims at a true conversion of the way we look at other
people in the ordinary circumstances of life” (Bourdieu, 1999c, p. 614).
153
4.4.4 Transcribing
Transcribing is the procedure of translation from oral conversations to written texts (Kvale,
1996). Raw interview data in this study consisted of audio recordings and field notes, which
were converted into transcripts. They were first organised and prepared for analysis by
transcribing audio recording and typing up field notes. The use of transcribed data enhances
the precision, and hence the reliability, of the qualitative analysis (Peräkylä, 2004; Silverman,
1993).
Reliable and valid transcriptions are consistent and accurate descriptions of the data.
The descriptive aspect of data is crucial so readers can make their own interpretations (Patton,
1990). Reliable and valid transcribing renders the interview conversations in a format
amenable to closer analysis.
Transcription reliability can be improved by listening to the audio recording a few
times. This increased the consistency of the transcripts. ‘Member checking’ was also used to
achieve transcription reliability. The supervisory team reviewed the transcripts and helped
identify some vague language that needed to be clarified. The researcher checked these
passages by listening to the audio recording again. Where the researcher struggled to capture
what exactly was said by the interviewees, native English speakers were invited to help with
the transcription. This ensured that the interviewees’ replies were transcribed as accurately as
possible.
Ascertaining transcription validity is more complex than assuring its reliability. Kvale
(1996) suggested the combination of several techniques to improve validity. Firstly, verbatim
descriptions respect the original meanings of the participants whose own words will be
transcribed as much as possible. The researcher of the current study transcribed the wording
of the participants as accurately as possible by listening to the audio recording many times.
Secondly, nonverbal communications were recorded on the field notes where necessary and
154
possible. These forms of communication, such as pauses, laughter, interruptions, and changes
in vocal tone or emotion, provided a richer access to the participants’ meanings. Thirdly,
transforming the conversation into a written form facilitates communication of the
participants’ meanings to readers. While transcribing, the researcher corrected superficial
conversational grammar mistakes of the participants for the sake of reading clarity. Care was
taken to preserve meaning. As oral language uses different clause patterns from written
language, the talk was broken up into sentences according to the researcher’s own
understanding and later checked by the participants. In brief, the combination of verbatim
descriptions, nonverbal communications, and written styles made the transcription more
meaningful and manipulable.
Kvale’s advice resonates with Bourdieu’s viewpoints that transcription is subject to the
reconciliation of two sets of constraints: the constraint of fidelity and “the constraint of
readability” (Bourdieu, 1999c, p. 622). Transcription means writing and rewriting. The
writing process is faithful to everything that came up in the interview, while the rewriting
process sometimes has to rid the transcribed text of certain confused phrases or verbal slips to
improve readability (Bourdieu, 1999c). Reconciliation of the two constraints is therefore
paradoxical. As Bourdieu (Bourdieu, 1999c, p. 622) explained, “the transition from the oral to
the written, with the changes in medium, imposes infidelities which are without doubt the
condition of a true fidelity.”
4.4.5 Analysing
Data analysis in qualitative studies is an ongoing process (Mertens, 2005). It does not occur
only at the end of the study as is typical in quantitative studies. Analysis in qualitative studies
is typically recursive (Mertens, 2005). By this is meant that findings are generated and
155
systematically built as successive pieces of data are produced. This analysis of the subjective
experience and the initial ‘objective’ analysis in the quantitative phase can fit within the same
methodological paradigm. Bourdieu (1981, p. 96) contended that objective analysis does not
contradict subjective analysis “of primary experience of the social world, and of the
immediate comprehension of the utterances, acts, or works of others”. It is necessary to pass
from the “statistical regularity or algebraic structure” to the principle of qualitative
investigation of subjective experience in the social world (Bourdieu, 1981, p. 94). The flow
of this qualitative approach to data analysis was administered as follows.
The first step was to obtain a general sense of the information by reading through all
the data. This initial review did not involve a careful reading for details. Instead, this initial
review was more comprehensive, examining all the different sources of data to understand
the scope of the data.
The next step was meaning condensation that entailed an abridgement of the meanings
expressed by the participants into shorter formulations (Kvale, 1996). The “natural meaning
units” (Kvale, 1996, p. 194), which were quite often long statements expressed by the
participants, were determined by the researcher. Long statements were compressed into
briefer statements in which the main sense of what was said was rephrased in a few words.
The main sense that dominated a natural meaning unit was stated as succinctly as possible.
The researcher attempted to read the participants’ main sense without prejudice or judgement
and to generalise participants’ main sense into briefer statements as understood by the
researcher. Meaning condensation thus involved a reduction of large interview texts into
briefer and more succinct formulations.
Coding was then conducted. It is the process of dividing the materials into manageable
chunks of meaning, identifying different segments, and labelling those categories with a term
that can be used across data sources (Lodico et al., 2006). A combination of predetermined,
156
theoretically informed, and emerging codes was used. Data were continually read, reread, and
re-examined. New codes were added as the data were reviewed. The initial codes were
gradually combined and reduced with the goal of eliminating overlap and producing a more
coherent view of the patterns in the data (Lodico et al., 2006). The software NVivo was used
to help with the coding process. This was time-effective compared to the traditional coding
on the hardcopy of the transcripts where texts were manually marked with different colours.
Categories for segments were described by rendering the information in detail. That
detailed information was used to generate a small number of essential and non-redundant
themes. These themes were described using a few words or phrases but they identified the
major issues to interpret and represent the data. These emergent themes served as major
findings in the qualitative phase and helped structure the discussion. They displayed
perspectives from participants by selected and representative quotations and interpretations.
By doing this, different participants’ experiences were reconstructed into a richer, more
condensed and coherent story than the scattered stories of the separate participants. More
sophisticated analysis was then able to proceed. That analysis went beyond information
description and theme identification to a search for theoretical connections.
4.4.6 Reporting
Reporting is the end product of a long analytical process in interview research. The main aim
of a report is to demonstrate the importance and the trustworthiness of the findings (Kvale,
1996). Reporting is not simply representing the views of the participants, accompanied by the
researcher’s viewpoints in the form of interpretations. Rather, it is a social construction
whereby the researcher’s voice, writing style, and literary devices construct a specific view
on the participants’ lived world (Kvale, 1996). What is worth reporting is to be conveyed in a
157
limited number of pages, presenting the main aims, methods, results, and implications of the
interview research, and is thus necessarily selective.
The common mode of presenting the findings of the interview research is through
selected quotes. Kvale (1996) suggested several strategies when reporting quotes. First of all,
quotes should be contextualised to render the interview context, which the research knows
well but which is unknown to the reader. Secondly, quotes should be interpreted by clear
statements of what viewpoint a quote illuminates, proves, or disproves. Moreover, quotes
should be rendered in a written style. Verbal transcriptions of oral speech, with repetitions,
digressions, and pauses, are difficult to grasp when presented in a written form. To facilitate
comprehension, the participants’ spontaneous oral speech should be rendered into a readable
and written textual form in the final report. The exception is when the linguistic form itself is
important to the study (Kvale, 1996). Last but not least, there is also an issue regarding the
balance between quotes and texts. The quotes should not make up more than half of the text
(Kvale, 1996) because many quotes with few connecting comments and interpretations can
appear chaotic and produce a linguistic flicker effect. Too many quotes mean that the data are
expected to speak for themselves. Instead, the researcher should have a theoretical frame to
inform the reader of the data by interpretation of the quotes.
The possibility of presenting the interview results in visual forms should not be
overlooked, although interview data are typically of a verbal nature. To this end, summary
tables were used in Chapter Six to report aspects of the interview results.
4.5 Ethical considerations
Ethics is not an afterthought or burden but an integral part of research design and
implementation. At all times, ethical considerations should be at the forefront of the
158
researcher’s agenda. Ethics functions as a form of guardianship against inappropriate research
behaviours, such as impropriety, deception, disturbance, offence, discrimination, and invasion
of privacy. Ethics also seeks to guarantee beneficence, respect, and justice in the research.
Therefore, sound ethics and sound methodology are mutually constitutive. Deeper concerns
about the rights, welfare, and betterment of participants tend to result in greater involvement
and engagement of the participants in the research process, and ultimately, richer data.
The current study included working and interacting with humans to conduct
questionnaires and interviews in Australia. This research was considered to pose a low risk to
participants. All participants were adults and the theme of the study was related to their
Chinese cultural heritage and CHL learning experiences. There was no intent to elicit
sensitive or intrusive data. Participation in the questionnaire or interview was informed and
voluntary. Participants were provided with detailed information about the study (see
Appendix 1 and 2). They could withdraw from the study if they felt uncomfortable when
doing the questionnaire or interview, which was an option but was not taken up by any
participant. The research was conducted in accordance with the National Statement on Ethical
Conduct in Research Involving Humans (2007) and the ethical guidelines of Queensland
University of Technology Human Ethics Committee. An application for ethical review was
submitted to the low risk review team in the Faculty of Education. The low risk application
was reviewed by the Faculty Research Ethics Advisor and confirmed as meeting the
requirements of the National Statement on Ethical Conduct in Human Research. The current
research has received approval (number 1100000165) from the Research Ethics Unit at
Queensland University of Technology.
Key risks related to the potential for distress if the questionnaire or interview questions
inadvertently elicited recall of unpleasant cultural or learning experiences, and to
inconvenience arising from the times, duration, and frequency of interviews. The researcher
159
carefully managed these potential risks. Consequently, no participants seemed to feel
uncomfortable about participating in the research.
An online questionnaire was used so that the privacy of participants was protected. The
data that participants provided were handled in such a way that no uniquely identifying
information was attached to the data and nobody could trace the data back to the participants
who provided them. There was an exception if the participants of the online questionnaire
were interested in participating in the interview afterwards. They were asked to leave their
contact details to arrange a comfortable time and place for interviews. The researcher thus
knew the identities of the participants in the interviews. However, an identity code was used
on any transcripts and pseudonyms were used in any reports. These participants’ identities
were only known and available to the researcher. They were not disclosed. Identification was
removed from the stored data. All paper-based documents were kept securely in a locked
filing cabinet in the researcher’s QUT office. E-files were stored on a QUT
password-protected network drive. USB drives were only used for data transfer. Only the
researcher had access to the raw data.
Ethics considerations promoted the integrity of the current research. However, ethical
considerations extended further than the promise of confidentiality and anonymity.
Participants volunteered to share their life stories. Therefore, they deserved the highest
standards of ethical treatment. The researcher endeavoured to protect the participants and
maintain their trust and respect. Not only participants but also research sites were respected.
The researcher displayed this respect by gaining permission before entering any site and by
being careful to minimise any disruption to the physical settings.
The whole process of this research paid great heed to ethical considerations. Research
design, data production, analysis and interpretation, final writing, and future dissemination of
the research all reflected cognisance of ethical practice.
160
4.6 Chapter summary
This chapter prepared the path for the research. The rich methodological pluralism wove the
initial quantitative phase and the subsequent qualitative phase together, helped think through
‘objectivism’ and ‘subjectivism’, and reconciled ‘agency’ and ‘structure’. As such, it
resonates with the Bourdieusian stances methodologically and theoretically. The selection of
an explanatory mixed methods design was based on philosophical, theoretical, and
methodological considerations. This approach established the basis for the detailed research
design that enabled data production, collection, and analysis to address the two research
questions. The two sub-theoretical frameworks, shown in Figure 4.1 and Figure 4.4, guided
the initial quantitative phase and the subsequent qualitative phase respectively, each of which
addressed one research question. The combination of quantitative and qualitative approaches
talked back to the overall theoretical framework and guided the data analysis and discussion.
Guided by the methodology chapter, the next two chapters report on the data analysis,
demonstrate the results, and reveal the findings. In the initial quantitative phase, SEM tests
the first part of the theoretical framework (Figure 4.1) by fitting the data into the statistical
models. The subsequent qualitative investigation is guided by the second part of the
theoretical framework (Figure 4.4). The detailed quantitative and qualitative data analysis and
results will be presented in the following Chapter Five and Chapter Six.
161
Chapter 5: Quantitative Data Analysis and Findings
This chapter reports the process of the quantitative analysis of the data. It also reports the
findings from the quantitative phase. The focus of this chapter addresses RQ1: Is CHL
proficiency of young Chinese Australian adults influenced by their investment of capital, the
strength of their habitus of ‘Chineseness’, or both? The first part of the theoretical framework
(Figure 4.1) guides the examination of this research question. Section 5.1 reports on the pilot
study, including issues related to face validity and internal consistency reliability of the
measurements. Section 5.2 details the quantitative data analysis of the main study and
explicates the results. EFA and CFA function to develop the measurement models for the six
theoretical constructs, namely ‘Chineseness’, ‘economic capital’, ‘cultural capital’, ‘social
capital’, ‘symbolic capital’, and ‘CHL proficiency’, specified as latent variables with their
corresponding indicators. They also function to validate the reliability and validity of the
measurement models. SEM constructs the relationship between participants’ ‘Chineseness’
and their ‘CHL proficiency’ as well as the relationship between their ‘capital’ and their ‘CHL
proficiency’, which serves the hypothesis testing. Multiple regression models function to
investigate whether participants’ habitus of ‘Chineseness’ and their investment of ‘capital’
co-contribute to their ‘CHL proficiency’. Section 5.3 highlights the key findings from the
quantitative data and concludes with a brief summary.
5.1 Pilot study
In the first stage of the pilot study, a printed copy of the questionnaire was distributed to five
participants, who were known to the researcher. Face validity and language appropriateness
were checked by virtue of the feedback from the participants. This ‘pre-pilot’ process was
162
important in that it identified some problems with the instrument. These problems and the
resultant modification of the instrument are discussed below.
In the original design, all participants were asked to choose which generation of
Chinese Australians they represented. A Malaysian-born Chinese Australian participant
commented that he considered all non-Australian-born Chinese Australians first-generation
Chinese Australians, no matter whether they moved to live in Australia by themselves or with
their parents. According to generation demarcation of the current study, participants born
outside Australia were treated as first-generation Chinese Australians. Therefore, they should
not have been asked about their generation. Only those born in Australia were asked this
question.
Participants were asked how many times they had visited China while living in
Australia. One participant had visited China while living in Singapore. She asked whether
this should count or not. The purpose of this item was to find out how frequently participants
had visited China. Therefore, this item was reworded as “How many times have you visited
China?”
Participants were also asked whether they had lived in a country other than Australia
and their country of birth. Some participants asked how long could be regarded as “have
lived”. This item was then clarified as “Have you ever lived in a country other than Australia
and your country of birth for more than one year?”
The above examples demonstrated how the pre-pilot process helped identify issues and
‘fine-tune’ the instrument. In this way, an online questionnaire was developed for the second
stage of the pilot study. Before reporting the second stage of the pilot study, the distinctions
between a reflective model and a formative model explained in Chapter Four were revisited.
The major statistical tests used in the current study are all reflective per se. These tests
included inter-item reliability test, Exploratory Factor Analysis (EFA), Confirmatory Factor
163
Analysis (CFA), and Structural Equation Modelling (SEM). Before conducting these tests,
the questionnaire items were revisited. Items inappropriate to plumb reflective models were
excluded from these tests. But some of the excluded items were still kept in the questionnaire
because they can capture meaningful information for the current study. For example, the item
(v10) measuring the individual gross annual income was not appropriate for reflective models
because the individual gross annual income may or may not vary with the variation of the
individual investment of economic capital in CHL learning. However, this item (v10) can
offer information about participants’ economic status. Therefore, it was kept in the online
questionnaire but was not included in the inter-item reliability test.
At the second stage of the pilot study, the link to the online questionnaire was sent to
38 participants who met the demographic requirements of the study. Given that these
participants were all known to the researcher, the return rate was 100%. It will be recalled
that there are six constructs as latent variables, namely ‘Chineseness’, ‘economic capital’,
‘cultural capital’, ‘social capital’, ‘symbolic capital’, and ‘CHL proficiency’. The online
questionnaire items were treated as indicator variables for these constructs. In the pilot study,
the item sets used to map the corresponding constructs are detailed in Table 5.1.
164
Table 5.1
Item sets and their corresponding constructs in the pilot study
Constructs Indicators Item description
Chineseness
v1 My mathematics was much better than that of my classmates.
v2 Handwriting tells you a lot about a person’s character.
v3 Children must be taught to paint, dance, or play musical instruments.
v4 Academic education is the most important thing in school.
v5 I prefer to live close to my family members.
v6 I hope to have sons to continue my family line or my husband’s family line.
v7 People who are senior in age and/or position should be addressed by their
title plus surname rather than their first name.
v8 To save face I always prefer to say ‘yes’.
v9 Confucianism values the virtue of frugality. Confucius said that frugality is
a common character of people with virtue. He also said that frugality is
very important in family life. In general, to what degree would you agree
with these values?
Economic
capital
v12 Money is not an issue if I wish to travel to China.
v13 I can afford to study in Chinese language schools as many hours as I wish.
v14 I can afford to employ private tutors to teach me Chinese if I wish.
v15 Money is never an issue if I want to buy Chinese language learning
materials, such as textbooks, dictionaries, and tapes.
Cultural
capital
v16 Learning Chinese is very important to me.
v17 I feel very sad if I miss celebrating Chinese festivals (Spring Festival,
Mid-Autumn, etc.).
v18 I always keep up to date with current Chinese affairs by watching TV,
listening to the radio, reading newspapers, or surfing online.
v19 I always go to venues, such as libraries, galleries, museums, theatres, or
concerts, if they feature Chinese culture.
v20 I read a lot of books about China.
v21 I have invested a lot of time in practising Chinese cultural activities, such as
learning musical instruments, calligraphy, or painting.
Social
capital
v24 Most of my friends are of Chinese descent.
v25 I am a member of social groups, such as a church or club, which mostly
include members of Chinese descent.
v26 I tend to mix exclusively with Chinese social groups.
v27 I think it is necessary to learn Chinese language in order to socialise with
my Chinese peers.
Symbolic
capital
v31 The wealth of my family is well-known in Chinese communities.
v32 My Chinese language competency is well regarded.
v33 Learning Chinese has increased my status in Chinese communities.
v34 People value my Chinese heritage.
v35 People consider me very popular among my Chinese peers.
CHL
proficiency
v36 I can read Chinese newspapers and magazines easily.
v37 I can read Chinese popular fiction easily.
v38 I can express my personal preferences and opinions in very clearly written
Chinese.
v39 I seldom make grammatical errors when writing in Chinese.
v40 I can easily understand Chinese conversations between people.
v41 I can easily understand Chinese language in the media, such as TV shows,
videos, and movies.
v42 I can talk about my life in detail in Chinese.
v43 I can handle complex situations in Chinese, such as banking, arguing,
purchasing a house or a car.
v44 I make appropriate use of popular words or idiomatic phrases common in
Chinese-speaking communities.
165
The value of Cronbach’s α corresponding to all six latent variables was above the
cutoff value of .80 (Kline, 1999). The value of corrected item total correlation corresponding
to all items was well above the cutoff value of .33, indicating that each item accounted for
more than 10% of the variance of the corresponding theoretical construct (R. Ho, 2006).The
overall internal consistency reliability of the instrument was considered to be reasonably high.
Table 5.2 summarises the results of internal consistency reliability tests for the six latent
variables.
Table 5.2
Summary results of internal consistency reliability test
Variables Chineseness Economic
capital
Cultural
capital
Social
capital
Symbolic
capital
CHL
proficiency
Cronbach’s α .91 .84 .91 .90 .87 .98
A close investigation of the inter-item correlation matrix identified that some indicators
mapping the construct ‘CHL proficiency’ were highly correlated (>.90), which could be
indicative of multicollinearity (Field, 2009). This was because the highly correlated items
were not distinguishable enough to stratify participants’ CHL proficiency into different levels.
For example, if participants could read Chinese newspapers and magazines easily, they were
likely to read Chinese popular fiction stories easily. Their CHL reading proficiency could not
be distinguished by these two items. Statistically, these two items were redundant and thus
were highly correlated. Based on extensive discussions with experienced Chinese language
teachers and CHL learners at various levels, the following two items were believed to be able
to distinguish participants’ CHL reading proficiency: (1) “I can read Chinese language
textbooks easily”; and (2) “I can read Chinese popular fiction stories easily”. Similar
strategies were applied to revise the indicators measuring CHL writing, listening, and
speaking proficiencies.
166
The distribution of the data points was also checked. For example, the data
corresponding to the item asking personal gross income was heavily positively skewed. There
was a larger concentration at the lower end of the scores. This indicated that the income
levels of the scale might have been too large. In order to better capture the spread of income,
the income levels within the scale were set to a lower level.
A 7-point bi-polar Likert scale was first used in the pilot study. The scale ranged from 1
(strongly disagree) to 7 (strongly agree). Compared to the bi-polar scale, the uni-polar scale
was later considered less mentally taxing because participants only have to consider one
attitude instead of balancing two opposing attitudes. As such, a uni-polar approach is
preferred wherever possible (Groves et al., 2004). Therefore, the original bi-polar scale was
changed to a 7-point uni-polar Likert-type scale in the main study. The scale ranged from 1
(not at all) to 7 (completely).
The pilot process shaped and cleaned the items that were used to measure their
corresponding constructs. Some of the items were reworded. Some of the items were
removed. Consequently, the items were renumbered. The item sets used in the main study to
map the corresponding constructs are detailed in Table 5.3.
Table 5.3
Item sets and their corresponding constructs in the main study
Constructs Indicators Item description
Chineseness
v1 My mathematics was much better than that of my classmates. To what extent
do you agree with this statement?
v2 Handwriting tells you a lot about a person’s character. To what extent do you
agree with this statement?
v3 Children must be taught to paint, dance, or play musical instruments. To
what extent do you agree with this statement?
v4 Academic education is the most important thing in school. To what extent do
you agree with this statement?
v5 I prefer to live close to my family members. To what extent do you agree
with this statement?
v6 I hope to have sons to continue my family line or my husband’s family line.
To what extent do you agree with this statement?
v7 People who are senior in age and/or position should be addressed by their
title plus surname rather than their first name. To what extent do you agree
167
with this statement?
v8 To save face I always prefer to say ‘yes’. To what extent do you agree with
this statement?
v9 Confucianism values the virtue of frugality. Confucius said that frugality is a
common character of people with virtue. He also said that frugality is very
important in family life. In general, how much degree would you agree with
these values?
Economic
capital
v11 Money is not an issue if I wish to travel to China. To what extent do you
agree with this statement?
v12 I can afford to study in Chinese language schools as many hours as I wish.
To what extent do you agree with this statement?
v13 I can afford to employ private tutors to teach me Chinese if I wish. To what
extent do you agree with this statement?
v14 Money is never an issue if I want to buy Chinese language learning
materials, such as textbooks, dictionaries, and tapes. To what extent do you
agree with this statement?
Cultural
capital
v15 Learning Chinese is very important to me. To what extent do you agree with
this statement?
v16 I feel very sad if I miss celebrating Chinese festivals (Spring Festival,
Mid-Autumn, etc.). To what extent do you agree with this statement?
v17 I always keep up to date with current Chinese affairs by watching TV,
listening to the radio, reading newspapers, or surfing online. To what extent
do you agree with this statement?
v18 I always go to venues, such as libraries, galleries, museums, theatres, or
concerts, if they feature Chinese culture. To what extent do you agree with
this statement?
v19 I read a lot of books about China. To what extent do you agree with this
statement?
v20 I have invested a lot of time in practising Chinese cultural activities, such as
learning musical instruments, calligraphy, or painting. To what extent do you
agree with this statement?
Social
capital
v23 Most of my friends are of Chinese descent. To what extent do you agree
with this statement?
v24 I am a member of social groups, such as a church or club, which mostly
include members of Chinese descent. To what extent do you agree with this
statement?
v25 I tend to mix exclusively with Chinese social groups. To what extent do you
agree with this statement?
v26 I think it is necessary to learn Chinese language in order to socialise with my
Chinese peers. To what extent do you agree with this statement?
Symbolic
capital
v30 The wealth of my family is well-known in Chinese communities. To what
extent do you agree with this statement?
v31 My Chinese language competency is well regarded. To what extent do you
agree with this statement?
v32 Learning Chinese has increased my status in Chinese communities. To what
extent do you agree with this statement?
v33 People value my Chinese heritage. To what extent do you agree with this
statement?
v34 People consider me very popular among my Chinese peers. To what extent
do you agree with this statement?
CHL
proficiency
v35 I can easily understand my family members and friends when they talk to
me in Chinese. To what extent do you agree with this statement?
v36 I can easily understand Chinese language in the media, such as TV shows,
videos, and movies. To what extent do you agree with this statement?
v37 I can easily handle complex situations in Chinese, such as banking, arguing,
purchasing a house or a car. To what extent do you agree with this
168
statement?
v38 I can read Chinese language textbooks easily. To what extent do you agree
with this statement?
v39 I can have deep discussion and exchange thinking with people in Chinese.
To what extent do you agree with this statement?
v40 I can read Chinese popular fiction stories easily. To what extent do you
agree with this statement?
v41 I can always write Chinese characters and Chinese words correctly. To what
extent do you agree with this statement?
v42 I can express my personal preferences and opinions in very clearly written
Chinese. To what extent do you agree with this statement?
5.2 Main study
Two hundred and thirty people responded to the modified online questionnaire. Because the
‘not allow to skip question’ option was applied in Key Survey, respondents had to answer all
questions. By virtue of this strategy, all 230 cases were data complete. Due to the clearly
delineated eligibility for participation, all the respondents met the demographic requirements
of the study. They all ranged in age between 18 and 35. The age distribution demonstrated a
good spread (see Figure 5.1). Forty seven point eight percent (47.8%) of the respondents
were males and 52.2% were females. This gender distribution is comparable to ABS 2011
Census data, which reported 46.4% and 53.6% of Chinese Australians are males and females
respectively. As shown in Figure 5.2, all the respondents were from the capital cities where
the overwhelming proportion of the Chinese Australian population resided (93.6% according
to ABS 2011 Census data). The only capital city not represented in the study was Hobart.
This should not impact on the demographic diversity of the sample because the Chinese
Australian population in Tasmania was very small (according to ABS 2011 Census, 0.65% of
the Tasmanian population had Chinese ancestry). An overwhelming proportion (83.1%) of the
respondents lived in Sydney, Melbourne, and Brisbane. This proportion is comparable to ABS
2011 Census data, which reported 84.1%. In this vein, the demographic features of the
sample are largely consistent with those of the population.
169
Figure 5.1. Age distribution of participants
Figure 5.2. Distribution of participants’ resident cities
One hundred and nineteen respondents (51.7%) were born in Australia. Ninety five
respondents (41.4%) were born in China, including Chinese Mainland, Hong Kong, Macau,
and Taiwan. Sixteen respondents (6.9%) were born in other countries, including Indonesia,
Malaysia, New Zealand, Singapore, and Vietnam (see Figure 5.3). All the 111 non-Australian
born respondents (48.3% of the total respondents) came to live in Australia before the age of
13. They were regarded as first-generation Chinese Australians. Those born in Australia were
asked to designate their generations. Seventy-three of them (31.7% of the total respondents)
170
considered themselves to be second generation, 31 (13.5%) third generation, and 15 (6.5%)
fourth generation or further removed (see Figure 5.4).
Figure 5.3. Composition of participants’ birthplace
Figure 5.4. Composition of participants’ generation
One hundred and twelve respondents (48.7%) spoke both Chinese and English at home.
Sixty-four respondents (27.8%) only spoke Chinese at home and fifty-one respondents
(22.2%) only spoke English at home. Three respondents (1.3%) spoke other languages or
other languages mixed with Chinese and English at home, such as Indonesian, Indonesian
mixed with English and Chinese, or Vietnamese mixed with English and Chinese (see Figure
5.5). Of those who used Chinese at home, 109 (61.9%) used Mandarin, 55 (31.3%) used
Cantonese, and 12 (6.8%) used a mixture of Mandarin and Cantonese, or a mixture of
Mandarin and Hokkien, or other Chinese dialects, such as Fujianese, Hakka, Shanghainese,
Sichuanese, and Teochew (see Figure 5.6).
119, 51.7% 95, 41.4%
16, 6.9%
born in Australia
born in China
111, 48.3%
73, 31.7%
31, 13.5%
15, 6.5%
first generation
second generation
thrid generation
fourth generation and further removed
171
Figure 5.5. Composition of participant’s language usage at home
Figure 5.6. Composition of Chinese dialects used at home
An overwhelming proportion (87%) of the respondents had formally studied Chinese in
schools, universities, Confucius Institutes, and/or community schools. However, their
investment of time spent in studying Chinese varied from less than one year to over 15 years,
and their CHL proficiency varied remarkably. Data analysis reported in this chapter is to
explain how much variance of their CHL proficiency can be attributed to their habitus of
‘Chineseness’ and their investment of different forms of capital.
Descriptive statistics were applied before the main tests. Results are reported in Table
5.4. The standard deviations of the indicators ranged from 1.63 to 2.09 on a 7-point scale.
This suggested adequate variance in the responses.
112, 48.7%
64, 27.8%
51, 22.2%
3, 1.3% both English and Chinese
Chinese only
English only
other languages or other languages
mixed with English and Chinese
109, 61.9%
55, 31.3%
12, 6.8%
Mandarin
Cantonese
other dialects or a mixture of Mandarin and a dialect
172
Table 5.4
Descriptive statistics of indicators
Constructs Indicators N Mean SD Skewness Kurtosis
Stats Stats SE Stats Stats SE Stats SE
Chineseness
v1 230 4.55 .12 1.80 -.35 .16 -.84 .32
v2 230 4.11 .12 1.80 -.08 .16 -.92 .32
v3 230 4.12 .12 1.75 -.06 .16 -.82 .32
v4 230 4.40 .12 1.84 -.29 .16 -.95 .32
v5 230 4.37 .11 1.68 -.15 .16 -.72 .32
v6 230 3.62 .12 1.89 .20 .16 -1.04 .32
v7 230 4.26 .12 1.81 -.14 .16 -.97 .32
v8 230 3.68 .12 1.88 .15 .16 -1.07 .32
v9 230 4.18 .11 1.63 -.08 .16 -.67 .32
Economic
capital
v11 230 3.96 .12 1.82 .01 .16 -.98 .32
v12 230 4.02 .12 1.88 .03 .16 -1.08 .32
v13 230 3.88 .12 1.89 .10 .16 -1.09 .32
v14 230 4.22 .12 1.85 -.12 .16 -1.01 .32
Cultural
capital
v15 230 4.69 .11 1.74 -.42 .16 -.72 .32
v16 230 4.20 .12 1.84 -.11 .16 -.98 .32
v17 230 3.67 .12 1.87 .18 .16 -1.03 .32
v18 230 3.54 .12 1.83 .23 .16 -.98 .32
v19 230 3.35 .13 1.95 .40 .16 -1.06 .32
v20 230 3.52 .13 1.96 .30 .16 -1.12 .32
Social capital
v23 230 4.19 .12 1.79 -.07 .16 -.92 .32
v24 230 3.74 .13 1.92 .12 .16 -1.10 .32
v25 230 3.53 .12 1.76 .23 .16 -.90 .32
v26 230 4.07 .13 1.91 -.05 .16 -1.10 .32
Symbolic
capital
v30 230 3.36 .12 1.80 .31 .16 -.94 .32
v31 230 3.87 .12 1.88 .05 .16 -1.08 .32
v32 230 3.59 .12 1.83 .19 .16 -1.00 .32
v33 230 4.29 .12 1.77 -.12 .16 -.90 .32
v34 230 4.12 .11 1.71 -.06 .16 -.76 .32
CHL
proficiency
v35 230 5.16 .12 1.76 -.58 .16 -.90 .32
v36 230 4.40 .14 2.07 -.25 .16 -1.27 .32
v37 230 4.31 .13 2.00 -.14 .16 -1.26 .32
v38 230 3.82 .14 2.07 .12 .16 -1.31 .32
v39 230 4.04 .14 2.09 -.03 .16 -1.35 .32
v40 230 3.13 .13 2.01 .59 .16 -.94 .32
v41 230 3.63 .13 1.94 .26 .16 -1.10 .32
v42 230 3.06 .13 1.92 .65 .16 -.75 .32
5.2.1 Validity and reliability of measurement models: EFA and CFA
As EFA is being used in a descriptive way to summarise relationships, assumptions in regards
to normal distribution may be relaxed as long as the deviation is not too large (Tabachnick &
Fidell, 2007). Where possible a single factor congeneric measurement model was then
constructed for each latent variable. CFA was then conducted to demonstrate the degree to
173
which the indicators were reflective of their corresponding constructs. It reflected how well
the indicators captured the essence of the corresponding latent variables. Multivariate
normality was checked where it was required for the particular analysis adopted. When
multivariate normality was violated, bootstrapping techniques were applied to minimise the
impact of the non-normal distribution of the data. It is usual in a measurement model to set
the scale of the latent variable by fixing the variance of the construct to 1 rather than the usual
practice of setting a factor loading to 1. In this way, standard error estimates can be generated
for all factor loadings, which is of interest in a measurement model. However, this may
generate incorrect standard errors when bootstrapping techniques are applied. Consequently,
the factor loading of the indicator with the highest factor loading in EFA was set to 1. The
sections below report on the process of EFA and CFA that shape the instrument and improve
the reliability and validity of the measurement models.
‘Chineseness’
It will be recalled that Chapter Three introduced a core set of Confucian values consisting of
Wuchang (五常, Five Constants), Sizi (四字, Four Virtues), and Liuyi (六艺, Six Skills).
These Confucian values serve as the cultural and historical foundation of the habitus of
‘Chineseness’. As argued in Chapter Four, this habitus of ‘Chineseness’ can be quantified by
operationalising key dimensions of Confucian norms, ideals, and social perspectives
documented in Confucian classics and popular classical sayings. Originally, nine indicators,
reflective of durable and transposable dispositions rooted in Confucianism, were used to
measure the habitus of ‘Chineseness’. This section discusses whether the nine indicators fit
into the measurement model for the construct ‘Chineseness’, then reports on the modification
of the measurement model to improve the model fit.
174
Reflective models require intercorrelation between indicators. As such, the correlation
matrix was examined in the first step of EFA. It is desirable that the inter-item correlations
should be greater than .30 to demonstrate enough intercorrelation but less than .90 to avoid
multicollinearity (Field, 2009). The correlation matrix for indicators corresponding to
‘Chineseness’ demonstrated that all the indicators had a statistically significant small to
medium correlation with other indicators (p < .001 as shown in Table 5.5), varying within the
cutoff range between .30 and .90.
Table 5.5
Correlation matrix for the indicators within the construct ‘Chineseness’
Correlation v1 v2 v3 v4 v5 v6 v7 v8 v9
v1 1.00 .54 .52 .52 .49 .41 .43 .50 .49
v2 .54 1.00 .57 .62 .44 .54 .61 .56 .50
v3 .52 .57 1.00 .55 .43 .51 .48 .47 .51
v4 .52 .62 .55 1.00 .47 .48 .60 .57 .52
v5 .49 .44 .43 .47 1.00 .37 .49 .45 .52
v6 .41 .54 .51 .48 .37 1.00 .58 .66 .47
v7 .43 .61 .48 .60 .49 .58 1.00 .59 .59
v8 .50 .56 .47 .57 .45 .66 .59 1.00 .54
v9 .49 .50 .51 .52 .52 .47 .59 .54 1.00
Sig. v1 v2 v3 v4 v5 v6 v7 v8 v9
v1 .000 .000 .000 .000 .000 .000 .000 .000
v2 .000 .000 .000 .000 .000 .000 .000 .000
v3 .000 .000 .000 .000 .000 .000 .000 .000
v4 .000 .000 .000 .000 .000 .000 .000 .000
v5 .000 .000 .000 .000 .000 .000 .000 .000
v6 .000 .000 .000 .000 .000 .000 .000 .000
v7 .000 .000 .000 .000 .000 .000 .000 .000
v8 .000 .000 .000 .000 .000 .000 .000 .000
v9 .000 .000 .000 .000 .000 .000 .000 .000
The Kaiser-Meyer-Olkin (KMO) measure of sampling adequacy is an index for
comparing the magnitude of the observed correlation coefficients to the magnitude of the
partial correlation coefficients. When the sum of the squared observed correlation coefficients
is remarkably larger than the sum of the squared partial correlation coefficients, the value of
KMO approaches 1. The higher the KMO value, the more compact the correlation of the
indicators. The KMO measure should be greater than .50 for a satisfactory factor analysis to
175
proceed (Kaiser, 1974). The KMO measure of sampling adequacy was .92, which
demonstrated that the EFA should yield distinct and reliable factors (see Table 5.6).
Bartlett’s test of sphericity is used to test the null hypothesis that there is no statistically
significant difference between the observed correlation matrix and the identity matrix where
the indicators in the correlation matrix are uncorrelated. The result of the Bartlett’s test of
sphericity rejected the null hypothesis and confirmed the adequacy of the magnitude of the
correlations by presenting a statistically significant chi-square value of 1028.82 (p < .001 as
shown in Table 5.6).
Table 5.6
KMO and Bartlett's test for the indicators within the construct ‘Chineseness’
KMO and Bartlett's test
Kaiser-Meyer-Olkin measure of sampling adequacy. .92
Bartlett's test of sphericity Approx. chi-square 1028.82
df 36
Sig. .000
An examination of communalities indicated that a reasonable level of variation of the
indicators (between 39% and 60% as highlighted in Table 5.7) was explained by the construct
‘Chineseness’. The initial communalities demonstrated how much variance the indicator had
in common with other indicators. The extraction communalities demonstrated how much
variance the indicator had in common with the extracted factors in comparison to the variance
explained by a linear combination of the other indicators. The communalities demonstrated
that the extraction in most cases explained more variance in an indicator than the initial
model. This indicated the appropriateness of the solution for the current factor extraction.
However, of concern were v1, v5, and v6, whose variances explained by the extraction were
not more than those explained by the initial model. These indicators will be examined more
closely subsequently.
176
Table 5.7
Communalities for the indicators within the construct ‘Chineseness’
Initial Extraction
v1 .44 .44
v2 .56 .60
v3 .47 .49
v4 .54 .58
v5 .39 .39
v6 .52 .50
v7 .57 .60
v8 .57 .58
v9 .48 .51
Extraction method: Maximum likelihood
Eigenvalues represent the amount of variation explained by a factor and an eigenvalue
of 1 or greater represents a substantial amount of variation (Kaiser, 1960). Of all the nine
impact factors for ‘Chineseness’, there was only one factor with an eigenvalue above 1.
Therefore, only one factor was extracted. This factor accounted for 57.18% of the total
variance of the nine indicators (see Table 5.8).
Table 5.8
Total variance of the construct ‘Chineseness’ explained by its indicators
Factor Initial eigenvalues Extraction sums of squared loadings
Total % of Variance Cumulative % Total % of Variance Cumulative %
1 5.15 57.18 57.18 4.67 51.92 51.92
2 .75 8.32 65.50
3 .63 6.98 72.48
4 .53 5.86 78.35
5 .50 5.52 83.87
6 .45 5.02 88.90
7 .39 4.34 93.24
8 .32 3.56 96.80
9 .29 3.20 100.00
Extraction method: Maximum likelihood
Examination of the scree plot also indicated the appropriateness of a one-factor
solution. The cutoff point for selecting the extracted factors should be at the point of inflexion
where the slope of the line changes dramatically (Cattell, 1966). The number of the extracted
factor(s) depends on the number of inflexion point(s). In Figure 5.7, the point of inflexion
177
occurred at the second data point, where there was a sharp descent in the curve followed by a
tailing off. Therefore, only one factor was extracted.
Figure 5.7. Scree plot of the ‘Chineseness’ scale
Before conducting CFA, univariate and multivariate normality were checked. As
highlighted in Table 5.9, all the individual indicators departed significantly from univariate
normal distribution by displaying the critical ratio of skewness, kurtosis, or both beyond the
cutoff value of |2| (Field, 2009). The value of Mardia’s coefficient was 16.14, which was
greater than the cutoff value of 3 (Yuan et al., 2002). This indicated significant deviation from
multivariate normality. Therefore, bootstrapping techniques were applied when conducting
CFA.
Table 5.9
Assessment of normality for the indicators within the construct ‘Chineseness’
Variable min max skew c.r. kurtosis c.r.
v9 1.00 7.00 -0.08 -0.52 -0.68 -2.10
v8 1.00 7.00 0.15 0.90 -1.07 -3.31
v7 1.00 7.00 -0.14 -0.87 -0.97 -3.01
v6 1.00 7.00 0.20 1.24 -1.04 -3.23
v5 1.00 7.00 -0.15 -0.90 -0.73 -2.27
v4 1.00 7.00 -0.29 -1.81 -0.95 -2.95
v3 1.00 7.00 -0.06 -0.37 -0.83 -2.56
v2 1.00 7.00 -0.08 -0.52 -0.93 -2.87
v1 1.00 7.00 -0.35 -2.18 -0.85 -2.62
Multivariate 16.14 8.70
178
A single-factor measurement model for the construct ‘Chineseness’ was specified as a
latent variable with nine indicators. The model with its standardised parameters is illustrated
in Figure 5.8.
Figure 5.8. Measurement model for the construct ‘Chineseness’
As highlighted in Table 5.10, the values of all the baseline comparisons fit indices (NFI,
RFI, IFI, TLI, and CFI) were above the cutoff value of .90 (Bentler, 1990), but the significant
chi-square value (CMIN = 66.91, p < .001) and the value of RMSEA (.08), which reached the
cutoff value of .08 (R. Ho, 2006), indicated that the model fit could be improved.
Table 5.10
Model fit indices for the construct ‘Chineseness’
CMIN
Model NPAR CMIN DF P CMIN/DF
Default model 18 66.91 27 .000 2.48 Saturated model 45 .000 0
Independence model 9 1046.34 36 .000 29.07 Baseline comparisons
Model NFI
Delta1 RFI
rho1 IFI
Delta2 TLI
rho2 CFI
Default model 0.94 0.92 0.96 0.95 0.96
Saturated model 1.00 1.00 1.00
Independence model 0.00 0.00 0.00 0.00 0.00
RMSEA
Model RMSEA LO 90 HI 90 PCLOSE
Default model 0.08 0.06 0.11 0.02
Independence model 0.35 0.33 0.37 0.00
179
An examination of the standardised residual covariance matrix (see Table 5.11) and the
modification index (see Table 5.12) indicated that the model fit can be improved by deleting
v1, v5, v6, and v8 from the model. These indicators were either associated with high values
of standardised residual covariances, or associated with high values of covariances between
indicator error terms or regression weights between indicators (as highlighted in Tables 5.11
and 5.12). It will be recalled that v1, v5, and v6 were also of concern in EFA. The
questionnaire items corresponding to these indicators were revisited. On reflection, these four
items did appear different from the others because they directly scrutinised the respondents’
personal viewpoints. In contrast, the other items scrutinised respondents’ viewpoints about a
general statement. By virtue of the reflective nature of the model, deleting indicators will not
necessarily change the conceptual domain of the theoretical construct (Coltman et al., 2008).
Therefore, these four indicators were tentatively excluded from the model.
Table 5.11
Standardised residual covariances for the indicators within the construct ‘Chineseness’
v9 v8 v7 v6 v5 v4 v3 v2 v1
v9 0.00
v8 -0.03 0.00
v7 0.46 0.08 0.00
v6 -0.45 1.61 0.47 0.00
v5 1.03 -0.35 0.20 -1.00 0.00
v4 -0.25 -0.12 0.16 -0.82 -0.03 0.00
v3 0.19 -0.75 -0.73 0.23 -0.01 0.29 0.00
v2 -0.63 -0.37 0.19 -0.05 -0.50 0.39 0.44 0.00
v1 0.17 -0.05 -1.15 -0.78 1.05 0.24 0.70 0.39 0.00
180
Table 5.12
Modification indices for the construct ‘Chineseness’
Covariances Regression weights
M.I. Par change
e6 <--> e8 19.36 0.53
e5 <--> e9 4.82 0.23
e5 <--> e6 4.51 -0.26
e4 <--> e6 5.02 -0.26
e3 <--> e8 4.12 -0.23
e3 <--> e7 4.14 -0.22
e1 <--> e7 9.17 -0.34
e1 <--> e5 4.29 0.26
Before testing the fit of the modified model, multivariate normality was checked again
because four indicators had been dropped from the model (see Table 5.13). The Mardia’s
coefficient value of 3.59, which was greater than the cutoff value of 3.00 (Yuan et al., 2002),
indicated the violation of multivariate normality. Therefore, the model was tested for fit by
using bootstrapping techniques.
Table 5.13
Assessment of normality for the indicators within the construct ‘Chineseness’ (modified)
Variable min max skew c.r. kurtosis c.r.
v9 1.00 7.00 -0.08 -0.52 -0.68 -2.10
v7 1.00 7.00 -0.14 -0.87 -0.97 -3.01
v4 1.00 7.00 -0.29 -1.81 -0.95 -2.95
v3 1.00 7.00 -0.06 -0.37 -0.83 -2.56
v2 1.00 7.00 -0.08 -0.52 -0.93 -2.87
Multivariate 3.59 3.25
A modified measurement model for the construct ‘Chineseness’ was specified as a
latent variable with five indicators. The modified model with standardised parameters is
illustrated in Figure 5.9.
M.I. Par change
v8 <--- v6 8.96 0.14
v7 <--- v1 4.80 -0.10
v6 <--- v8 7.28 0.13
181
Figure 5.9. Modified measurement model for the construct ‘Chineseness’
The modified model fitted the data well (see Table 5.14). It presented a non-significant
chi-square value of 10.54 (p = .061). All values for baseline comparison indices were above
the cutoff .90 (Bentler, 1990) and RMSEA value of .07 was below the cutoff .08 (R. Ho,
2006). The unstandardised regression weight of each indicator corresponded to a significant
critical ratio. All the values of the standardised regression weight (see Figure 5.9) were above
the cutoff value of .50 (Hair et al., 2006) and were above or the same as the preferred value
of .70 (Hair et al., 2006). This indicated that all five items were significant reflective
indicators of the corresponding construct ‘Chineseness’. The model converged.
Table 5.14
Model fit indices for the construct ‘Chineseness’ (modified)
CMIN
Model NPAR CMIN DF P CMIN/DF
Default model 10 10.54 5 .061 2.11
Saturated model 15 0.00 0
Independence model 5 494.08 10 .000 49.41
Baseline comparisons
Model NFI
Delta1 RFI
rho1 IFI
Delta2 TLI
rho2 CFI
Default model 0.98 0.96 0.99 0.98 0.99
Saturated model 1.00 1.00 1.00
Independence model 0.00 0.00 0.00 0.00 0.00
RMSEA
Model RMSEA LO 90 HI 90 PCLOSE
Default model .07 .000 .13 .24 Independence model .46 .43 .50 .00
182
The modified model had a significantly better fit than the original model. The modified
model had a chi-square difference of 56.37 (66.91-10.54) from the original model. With a
difference in degrees of freedom of 22 (27-5), this chi-square difference was statistically
significant (p < .001).
The factor score weight and the proportional factor score weight are reported in Table
5.15. The proportional factor score weight of each indicator was calculated as the factor score
weight of each indicator divided by the sum of the factor score weight (Rowe, 2002).
Table 5.15
Proportional factor score weights for the indicators within the construct ‘Chineseness’
v9 v7 v4 v3 v2 Sum
Factor score weight .17 .20 .20 .15 .22 .94
Proportional factor score weight .18 .21 .21 .16 .23 .99
Taking account of individual and joint measurement error, the scale score for the
construct can be computed as a continuous variable by multiplying the individual’s raw score
on each indicator by the proportionally weighted factor score of each indicator and summing
(Rowe, 2002). The scale score for the construct ‘Chineseness’ then can be calculated by the
following equation:
v9 × .18 + v7 × .21 + v4 × .21 + v3 × .16 + v2 × .23
This approach ensures that the estimates of the scale score adjusted for measurement
error are proportionally weighted by the actual contribution made by each indicator (Rowe,
2002). The proportional factor score sums to 1 hence the composite score will range from a
minimum of 1 to a maximum of 7. This process ensures that the construct has the same
‘metric’ as that of the indicators for the construct (Rowe, 2002). In Table 5.15, the
proportional factor score weights amounted to .99 due to rounding errors.
183
Reliability and validity of the construct ‘Chineseness’ were further checked. The
following sections reported four measures of reliability, namely Squared Multiple
Correlations (SMCs), construct reliability, variance extracted, and coefficient H, as well as
two measures of validity, namely construct validity and convergent validity.
The SMC for an indicator represents the proportion of variance in the indicator that is
explained by the construct. The SMC for an indicator greater than .30 is considered
acceptable while the SMC greater than .50 is preferred for a construct to have a good
mapping of that indicator (Jöreskog & Sörbom, 1996). As demonstrated in Figure 5.9, the
SMCs for all the five indicators were above the cutoff value of .30 and most of them were
above the preferred value of .50. The exceptions were the values of .49 and .48, which were
only marginally lower than .50. Therefore, the construct ‘Chineseness’ accounted for a
reasonable level of the variance of its indicators. The reliability of all the indicators of the
construct ‘Chineseness’ represented by the SMC measures was good.
The construct reliability compares the variance of the indicators captured by the
construct with that due to the measurement errors. The computation of the construct
reliability was offered by Fornell and Larcker (1981) as follows:
2 2
1 1 1
/ ( )n n n
i i i
i i i
r r e
In the formula above, n represents the number of indicators, r represents the standardised
regression weight of the indicator, and e represents the measurement error associated with the
indicator. The cutoff value for the construct reliability is .50 (Fornell & Larcker, 1981), which
indicates that the variance of the indicators captured by the construct exactly equals that due
to the measurement errors. As such, the greater the value is larger than .50, the better the
construct reliability. The construct reliability for ‘Chineseness’ was .56, which indicated that
the variance of the indicators captured by the construct ‘Chineseness’ was greater than that
due to the measurement errors. The construct reliability was acceptable.
184
The variance extracted expresses the overall amount of variance in the indicators
accounted for by the construct. Variance extracted exceeding 50% is commonly considered a
good measure of reliability (Fornell & Larcker, 1981). The construct ‘Chineseness’ accounted
for 64.56% of the overall variation in the indicators. Therefore, the reliability of the indicators
of the construct ‘Chineseness’ represented by the variance extracted was good.
Coefficient H (G. R. Hancock & Mueller, 2001) is becoming a popular way of
reporting reliability for SEM. The computation of coefficient H was offered by G. R.
Hancock and Mueller (2001) as follows:
2
21
1/ (1 1/ )1
ni
i i
r
r
In the formula above, n represents the number of indicators and r represents the standardised
regression weight of the indicator. A coefficient H value above .70 is desirable to achieve
reasonable reliability (G. R. Hancock & Mueller, 2001). The coefficient H value for the
construct ‘Chineseness’ was .87, which indicated a high reliability.
Construct validity requires the unidimensionality of the indicators. The model fit
measures can be viewed as confirming construct validity. The modified measurement model
for the construct ‘Chineseness’ had a good model fit, which supported the claim for its
construct validity.
Convergent validity, operationalised through the factor loading, is a measure of the
direct structural relationship between an indicator and the construct. To achieve convergent
validity, the factor loadings must be significantly different from 0. The critical ratios of the
unstandardised regression weights of the indicators are used to test this significance. As stated
earlier, the unstandardised regression weight of all the indicators of the construct
‘Chineseness’ corresponded to a significant critical ratio. Therefore, all the indicators
demonstrated a good convergent validity.
185
Economic capital
Four indicators were used to operationalise the construct ‘economic capital’. These four
indicators were reflective of the investment of financial resources in (potential) activities
associated with CHL learning. The correlation matrix demonstrated that all the indicators had
a statistically significant medium to strong correlation with other indicators (p < .001 as
shown in Table 5.16), varying within the cutoff range between .30 and .90 (Field, 2009). This
suggested the basis for EFA.
Table 5.16
Correlation matrix for the indicators within the construct ‘economic capital’
Correlation v11 v12 v13 v14
v11 1.00 .63 .59 .64
v12 .63 1.00 .80 .72
v13 .59 .80 1.00 .72
v14 .64 .72 .72 1.00
Sig. v11 v12 v13 v14
v11 .000 .000 .000
v12 .000 .000 .000
v13 .000 .000 .000
v14 .000 .000 .000
As shown in Table 5.17, the KMO measure of sampling adequacy (.83) was well above
the cutoff value of .50 (Kaiser, 1974), which demonstrated the EFA should yield distinct and
reliable factors. Bartlett’s test of sphericity confirmed the adequacy of the magnitude of the
correlations by presenting a statistically significant chi-square value of 573.18 (p < .001).
Table 5.17
KMO and Bartlett's test for the indicators within the construct ‘economic capital’
KMO and Bartlett's test
Kaiser-Meyer-Olkin measure of sampling adequacy. .83
Bartlett's test of sphericity Approx. chi-square 573.18
df 6
Sig. .000
186
An examination of communalities indicated that a reasonable level of variation of the
indicators (between 50% and 80% as highlighted in Table 5.18) was explained by the
construct ‘economic capital’. Specifically, the extraction in all cases explained more variance
in an indicator than the initial model (see Table 5.18). This indicated the appropriateness of
the solution for the current factor extraction.
Table 5.18
Communalities for the indicators within the construct ‘economic capital’
Initial Extraction
v11 .47 .50
v12 .70 .80
v13 .69 .78
v14 .62 .68
Extraction method: Maximum likelihood
Of all the four impact factors for ‘economic capital’, there was only one factor with an
eigenvalue above 1. Consequently, only one factor was extracted. This factor accounted for
76.46% of the total variance of the four indicators (see Table 5.19). Examination of the scree
plot also indicated the appropriateness of a one-factor solution (see Figure 5.10).
Table 5.19
Total variance of the construct ‘economic capital’ explained by its indicators
Factor Initial eigenvalues Extraction sums of squared loadings
Total % of Variance Cumulative % Total % of Variance Cumulative %
1 3.06 76.46 76.46 2.76 68.95 68.95
2 .45 11.28 87.74
3 .29 7.30 95.04
4 .20 4.96 100.00
Extraction method: Maximum likelihood
187
Figure 5.10. Scree plot of the ‘economic capital’ scale
As highlighted in Table 5.20, all the individual indicators departed significantly from
univariate normal distribution by displaying the critical ratio of kurtosis beyond the cutoff
value of |2| (Field, 2009). Mardia’s coefficient value of 8.21was greater than the cutoff value
of 3.00 (Yuan et al., 2002). This indicated significant deviation from multivariate normality.
Therefore, bootstrapping techniques were applied when conducting CFA.
Table 5.20
Assessment of normality for the indicators within the construct ‘economic capital’
Variable min max skew c.r. kurtosis c.r.
v14 1.00 7.00 -0.12 -0.73 -1.02 -3.14
v13 1.00 7.00 0.10 0.59 -1.10 -3.39
v12 1.00 7.00 0.03 0.20 -1.08 -3.35
v11 1.00 7.00 0.01 0.05 -0.98 -3.04
Multivariate 8.21 8.98
A single-factor measurement model for the construct ‘economic capital’ was specified
as a latent variable with four indicators. The model with standardised parameters is illustrated
in Figure 5.11.
188
Figure 5.11. Measurement model for the construct ‘economic capital’
Although the model had a significant chi-square value of 8.01 (p = .018), the ratio of
chi-square value and degree of freedom (4.00) was not very high. Although the RMSEA
value of .12 was above the cutoff value of .08 (R. Ho, 2006), all the baseline comparison fit
indices were above the cutoff value of .90 (Bentler, 1990). In particular, CFI and TLI, the two
most widely referenced indices (Hair et al., 2006), had relatively high values of .99 and .97
respectively (see Table 5.21). The unstandardised regression weight of each indicator
corresponded to a significant critical ratio. All the values of the standardised regression
weight (see Figure 5.11) were above the cutoff value of .50 (Hair et al., 2006) and the
preferred value of .70 (Hair et al., 2006). This indicated that all the four corresponding items
were significant reflective indicators of the respective construct ‘economic capital’. Therefore,
it can be argued that the model had an acceptable fit.
189
Table 5.21
Model fit indices for the construct ‘economic capital’
CMIN
Model NPAR CMIN DF P CMIN/DF
Default model 8 8.01 2 .018 4.00
Saturated model 10 .000 0
Independence model 4 578.65 6 .000 96.44 Baseline comparisons
Model NFI
Delta1 RFI
rho1 IFI
Delta2 TLI
rho2 CFI
Default model 0.99 0.96 0.99 0.97 0.99
Saturated model 1.00 1.00 1.00
Independence model 0.00 0.00 0.00 0.00 0.00
RMSEA
Model RMSEA LO 90 HI 90 PCLOSE
Default model 0.12 0.04 0.20 0.07
Independence model 0.65 0.60 0.69 0.00
The proportional factor score weight of each indicator is demonstrated in Table 5.22.
Table 5.22
Proportional factor score weights for the indicators within the construct ‘economic capital’
v14 v13 v12 v11 Sum
Factor score weight .20 .30 .34 .11 .95
Proportional factor score weight .21 .31 .36 .12 1.00
The scale score for the construct ‘economic capital’ can then be calculated by the
following equation:
v14 × .21 + v13 × .31 + v12 × .36 + v11 × .12
Four measures of reliability, namely SMC, construct reliability, variance extracted, and
coefficient H, as well as two measures of validity, namely construct validity and convergent
validity, are reported below.
As demonstrated in Figure 5.11, the SMCs for all the four indicators were above the
cutoff value of .30 (Jöreskog & Sörbom, 1996) and were the same as, or above, the preferred
value of .50 (Jöreskog & Sörbom, 1996). Therefore, the construct ‘economic capital’
accounted for a reasonable level of the variance of its indicators. The reliability of all the
190
indicators of the construct ‘economic capital’ represented by the SMC measures was
considered good. The construct reliability for ‘economic capital’ was .69, which was greater
than the cutoff value of .50 (Fornell & Larcker, 1981). The variance of the indicators captured
by the construct ‘economic capital’ was greater than that due to the measurement errors. This
indicated that the construct reliability was acceptable. The construct ‘economic capital’
accounted for 76.46% of the overall variation in the four indicators. This was well above the
cutoff of 50% (Fornell & Larcker, 1981). Therefore, the reliability of the indicators of the
construct ‘economic capital’ represented by the variance extracted was good. The coefficient
H value for the construct ‘economic capital’ was .91, which was well above the cutoff value
of .70 (G. R. Hancock & Mueller, 2001). This represented a high reliability.
All the four indicators of the construct ‘economic capital’ loaded on one factor, which
indicated the unidimensionality of the four indicators. This supported the claim for its
construct validity. As stated earlier, the unstandardised regression weight of all the indicators
of the construct ‘economic capital’ corresponded to a significant critical ratio. Therefore, all
the indicators demonstrated a good convergent validity.
Cultural capital
Originally, six indicators were used to operationalise the construct ‘cultural capital’. These
indicators were reflective of the cultural dispositions in relation to CHL learning and ‘taste’
associated with various Chinese cultural activities. This section discusses whether the six
indicators fit into the measurement model for the construct ‘cultural capital’ and reports on
the modification of the measurement model to improve the model fit.
The correlation matrix demonstrated that all indicators had a statistically significant
medium correlation with other indicators (p < .001 as shown in Table 5.23), varying within
191
the cutoff range between .30 and .90 (Field, 2009). The strength of the intercorrelation was
considered enough, without the problem of multicollinearity. This suggested the basis for
EFA.
Table 5.23
Correlation matrix for the indicators within the construct ‘cultural capital’
Correlation v15 v16 v17 v18 v19 v20
v15 1.00 .61 .51 .45 .54 .56
v16 .61 1.00 .66 .63 .57 .63
v17 .51 .66 1.00 .67 .65 .55
v18 .45 .63 .67 1.00 .72 .66
v19 .54 .57 .65 .72 1.00 .66
v20 .56 .63 .55 .66 .66 1.00
Sig. v15 v16 v17 v18 v19 v20
v15 .000 .000 .000 .000 .000
v16 .000 .000 .000 .000 .000
v17 .000 .000 .000 .000 .000
v18 .000 .000 .000 .000 .000
v19 .000 .000 .000 .000 .000
v20 .000 .000 .000 .000 .000
As shown in Table 5.24, the KMO measure of sampling adequacy (.87) was well above
the cutoff value of .50 (Kaiser, 1974), which demonstrated that the EFA should yield distinct
and reliable factors. Bartlett’s test of sphericity verified the adequacy of the magnitude of the
correlations by presenting a statistically significant chi-square value of 807.66 (p < .001).
Table 5.24
KMO and Bartlett's test for the indicators within the construct ‘cultural capital’
KMO and Bartlett's test
Kaiser-Meyer-Olkin measure of sampling adequacy. .87
Bartlett's test of sphericity Approx. chi-square 807.66
df 15
Sig. .000
An examination of communalities indicated that a reasonable level of variation in the
indicators (between 43% and 69% as highlighted in Table 5.25) was explained by the
192
construct ‘cultural capital’. The communalities in Table 5.25 demonstrated that the extraction
in most cases explained more variance in an indicator than the initial model. This indicated
the appropriateness of the solution for the current factor extraction. However, of concern was
v15, whose variance explained by the extraction was lower than that explained by the initial
model. This indicator will be examined more closely subsequently.
Table 5.25
Communalities for the indicators within the construct ‘cultural capital’
Initial Extraction
v15 .46 .43
v16 .60 .61
v17 .58 .62
v18 .65 .69
v19 .62 .67
v20 .58 .62
Extraction method: Maximum likelihood
There was only one factor with an eigenvalue above 1, accounting for 67.19% of the
total variance of the six indicators (see Table 5.26). Consequently, only one factor was
extracted. The scree plot also supported a one-factor solution (see Figure 5.12).
Table 5.26
Total variance of the construct ‘cultural capital’ explained by its indicators
Factor Initial eigenvalues Extraction sums of squared loadings
Total % of Variance Cumulative % Total % of Variance Cumulative %
1 4.03 67.19 67.19 3.64 60.74 60.74
2 .62 10.27 77.46
3 .46 7.65 85.11
4 .39 6.48 91.59
5 .27 4.43 96.03
6 .24 3.97 100.00
Extraction method: Maximum likelihood
193
Figure 5.12. Scree plot of the ‘cultural capital’ scale
As highlighted in Table 5.27, all the individual indicators departed significantly from
univariate normal distribution by displaying the critical ratio of skewness, kurtosis, or both
beyond the cutoff value of |2| (Field, 2009). Mardia’s coefficient value (8.26) was higher than
the cutoff value of 3.00 (Yuan et al., 2002), indicating significant deviation from multivariate
normality. Therefore, bootstrapping techniques were applied when conducting CFA.
Table 5.27
Assessment of normality for the indicators within the construct ‘cultural capital’
Variable min max skew c.r. kurtosis c.r.
v20 1.00 7.00 0.30 1.84 -1.12 -3.46
v19 1.00 7.00 0.39 2.44 -1.07 -3.30
v18 1.00 7.00 0.23 1.40 -0.98 -3.04
v17 1.00 7.00 0.18 1.12 -1.03 -3.19
v16 1.00 7.00 -0.11 -0.67 -0.99 -3.05
v15 1.00 7.00 -0.42 -2.58 -0.73 -2.27
Multivariate 8.26 6.39
A single-factor measurement model for the construct ‘cultural capital’ was specified as
a latent variable with six indicators. The model with standardised parameters is illustrated in
Figure 5.13.
194
Figure 5.13. Measurement model for the construct ‘cultural capital’
The model had a significant chi-square value of 52.48. The value of RFI (.89) did not
reach the cutoff value of .90 (Bentler, 1990). The value of RMSEA (.15) was much higher
than the cutoff value of .08 (R. Ho, 2006). These all indicated that the model did not fit the
data well (see Table 5.28).
Table 5.28
Model fit indices for the construct ‘cultural capital’
CMIN
Model NPAR CMIN DF P CMIN/DF
Default model 12 52.48 9 .000 5.83 Saturated model 21 .000 0
Independence model 6 817.78 15 .000 54.52 Baseline comparisons
Model NFI
Delta1 RFI
rho1 IFI
Delta2 TLI
rho2 CFI
Default model .94 .89 .95 .91 .95
Saturated model 1.00 1.00 1.00
Independence model .00 .00 .00 .00 .00 RMSEA
Model RMSEA LO 90 HI 90 PCLOSE
Default model .15 .11 .18 .00 Independence model .48 .46 .51 .00
An examination of the standardised residual covariance matrix (see Table 5.29) and the
modification index (see Table 5.30) indicated the model fit could be improved by deleting
v15 and v16 from the model. These two indicators were associated with a high value of
standardised residual covariances, a high value of covariances between indicator error terms,
and high values of regression weights between indicators (as highlighted in Table 5.29 and
195
5.30). It will be recalled that v15 was also of concern in EFA. Therefore, v15 and v16 were
tentatively excluded from the model.
Table 5.29
Standardised residual covariances for the indicators within the construct ‘cultural capital’
v20 v19 v18 v17 v16 v15
v20 0.00
v19 0.17 0.00
v18 0.11 0.52 0.00
v17 -0.87 0.04 0.25 0.00
v16 0.24 -0.92 -0.29 0.55 0.00
v15 0.55 -0.01 -1.34 -0.07 1.29 0.00
Table 5.30
Modification indices for the construct ‘cultural capital’
Covariances Regression weight
M.I. Par change M.I. Par change
e18<-->e19 5.64 0.22 v19<---v16 4.44 -0.09
e17<-->e20 9.81 -0.33 v18<---v15 9.28 -0.13
e16<-->e19 12.84 -0.36 v16<---v15 6.36 0.12
e15<-->e18 17.31 -0.42 v15<---v18 4.41 -0.10
e15<-->e16 11.93 0.38 v15<---v16 4.02 0.10
Before testing the fit of the modified model, multivariate normality was checked.
Mardia’s coefficient was 3.60, which was greater than the cutoff value of 3.00 (Yuan et al.,
2002). This implied the violation of the multivariate normality (see Table 5.31). Therefore,
bootstrapping was applied when conducting CFA.
Table 5.31
Assessment of normality for the indicators within the construct ‘cultural capital’ (modified)
Variable min max skew c.r. kurtosis c.r.
v20 1.00 7.00 0.30 1.84 -1.12 -3.46
v19 1.00 7.00 0.39 2.44 -1.07 -3.30
v18 1.00 7.00 0.23 1.40 -0.98 -3.04
v17 1.00 7.00 0.18 1.12 -1.03 -3.19
Multivariate 3.60 3.95
196
A modified measurement model for the construct ‘cultural capital’ was specified as a
latent variable with four indicators. The model with standardised parameters is illustrated in
Figure 5.14.
Figure 5.14. Modified measurement model for the construct ‘cultural capital’
The modified model had a non-significant chi-square value of 2.24 (p = .34). All the
baseline comparisons indices (NFI, RFI, IFI, TLI, and CFI) were well above the cutoff value
of .90 (Bentler, 1990). The value of RMSEA was .02, which was far below the cutoff value
of .08 (R. Ho, 2006). These implied a good model fit (see Table 5.32). The unstandardised
regression weight of each indicator corresponded to a significant critical ratio. All the values
of the standardised regression weight (see Figure 5.14) were above the cutoff value of .50
(Hair et al., 2006) and the preferred value of .70 (Hair et al., 2006). This indicated that all the
four corresponding items were significant reflective indicators of the respective construct
‘cultural capital’. The model converged.
197
Table 5.32
Model fit indices for the construct ‘cultural capital’ (modified)
CMIN
Model NPAR CMIN DF P CMIN/DF
Default model 8 2.24 2 .34 1.12 Saturated model 10 .00 0
Independence model 4 494.23 6 .00 82.37 Baseline comparisons
Model NFI
Delta1 RFI
rho1 IFI
Delta2 TLI
rho2 CFI
Default model 1.00 0.99 1.00 1.00 1.00
Saturated model 1.00 1.00 1.00
Independence model 0.00 0.00 0.00 0.00 0.00
RMSEA
Model RMSEA LO 90 HI 90 PCLOSE
Default model 0.02 0.00 0.14 0.51
Independence model 0.60 0.55 0.64 0.00
Compared to the original model, the modified model had a significantly better fit. The
modified model decreased the chi-square value of 50.24 (52.48-2.24) from the original model.
With the degree of freedom difference of 7 (9-2), the chi-square difference was statistically
significant (p < .001).
The proportional factor score weight of each indicator is demonstrated in Table 5.33.
Table 5.33
Proportional factor score weights for the indicators within the construct ‘cultural capital’
v20 v19 v18 v17 Sum
Factor score weight .16 .26 .32 .17 .91
Proportional factor score weight .18 .29 .35 .19 1.01
The scale score for the construct ‘cultural capital’ then can be calculated by the
following equation:
v20 × .18 + v19 × .29 + v18 × .35 + v17 × .19
Reliability and validity of the construct ‘cultural capital’ were further checked by four
measures of reliability, namely SMC, construct reliability, variance extracted, and coefficient
198
H, as well as two measures of validity, namely construct validity and convergent validity.
These measures are reported below.
As demonstrated in Figure 5.14, the SMCs for all the four indicators were above the
cutoff value of .30 (Jöreskog & Sörbom, 1996) and the preferred value of .50 (Jöreskog &
Sörbom, 1996). Therefore, the construct ‘cultural capital’ accounted for a reasonable level of
the variance of its indicators. The reliability of all the indicators of the construct ‘cultural
capital’ represented by the SMC measures was good. The construct reliability for ‘cultural
capital’ was .66, which was greater than the cutoff value of .50 (Fornell & Larcker, 1981).
The variance of the indicators captured by the construct ‘cultural capital’ was greater than
that due to the measurement errors. This indicated that the construct reliability was acceptable.
The construct ‘cultural capital’ accounted for 74.02% of the overall variation in the four
indicators. This was well above the cutoff of 50% (Fornell & Larcker, 1981). Therefore, the
reliability of the indicators of the construct ‘cultural capital’ represented by the variance
extracted was good. The coefficient H value for the construct ‘cultural capital’ was .89, which
was well above the cutoff value of .70 (G. R. Hancock & Mueller, 2001). This represented a
high reliability.
The modified measurement model for the construct ‘cultural capital’ had a good model
fit, which indicated the unidimensionality of the four indicators. This supported the claim for
its construct validity. As stated earlier, the unstandardised regression weight of all the
indicators of the construct ‘cultural capital’ corresponded to a significant critical ratio.
Therefore, all the indicators demonstrated a good convergent validity.
199
Social capital
Four indicators were used to operationalise the construct ‘social capital’. These indicators
were reflective of the size and strength of the social network consisting of members of
Chinese descent. This section discusses whether the four indicators fit into the measurement
model for the construct ‘social capital’.
The correlation matrix demonstrated that all the indicators had a statistically significant
medium correlation with other indicators (p < .001 as shown in Table 5.34), varying within
the cutoff range between .30 and .90 (Field, 2009). This suggested the basis for EFA.
Table 5.34
Correlation matrix for the indicators within the construct ‘social capital’
Correlation v23 v24 v25 v26
v23 1.00 .63 .68 .59
v24 .63 1.00 .59 .54
v25 .68 .59 1.00 .55
v26 .59 .54 .55 1.00
Sig. v23 v24 v25 v26
v23 .000 .000 .000
v24 .000 .000 .000
v25 .000 .000 .000
v26 .000 .000 .000
As shown in Table 5.35, the KMO measure of sampling adequacy (.82) was well above
the cutoff value of .50 (Kaiser, 1974), which demonstrated the EFA should yield distinct and
reliable factors. Bartlett’s test of sphericity confirmed the adequacy of the magnitude of the
correlations by presenting a statistically significant chi-square value of 399.57 (p < .001).
Table 5.35
KMO and Bartlett's test for the indicators within the construct ‘social capital’
Kaiser-Meyer-Olkin measure of sampling adequacy. .82
Bartlett's test of sphericity Approx. chi-square 399.57
df 6
Sig. .000
200
An investigation of communalities indicated that a reasonable level of variance in the
indicators (between 50% and 72% as highlighted in Table 5.36) was explained by the
construct ‘social capital’. The extraction in all cases explained more variance in an indicator
than the initial model (see Table 5.36). This indicated the appropriateness of the solution for
the current factor extraction.
Table 5.36
Communalities for the indicators within the construct ‘social capital’
Initial Extraction
v23 .57 .72
v24 .48 .57
v25 .52 .63
v26 .42 .50
Extraction method: Maximum likelihood
There was only one factor with an eigenvalue above 1, accounting for 69.99% of the
total variance of the four indicators (see Table 5.37). Consequently, only one factor was
extracted. The scree plot also supported a one-factor solution (see Figure 5.15).
Table 5.37
Total variance of the construct ‘social capital’ explained by its indicators
Factor Initial eigenvalues Extraction sums of squared loadings
Total % of Variance Cumulative % Total % of Variance Cumulative %
1 2.80 69.99 69.99 2.41 60.25 60.25
2 .47 11.84 81.83
3 .41 10.32 92.16
4 .31 7.84 100.00
Extraction method: Maximum likelihood
201
Figure 5.15. Scree plot of the ‘social capital’ scale
As highlighted in Table 5.38, all the individual indicators departed significantly from
univariate normal distribution by displaying the critical ratio of kurtosis beyond the cutoff
value of |2| (Field, 2009). The value of Mardia’s coefficient was 4.75, which was higher than
the suggested value of 3.00 (Yuan et al., 2002). This indicated that the multivariate normality
was violated. Therefore, bootstrapping techniques were applied when conducting CFA.
Table 5.38
Assessment of normality for the indicators within the construct ‘social capital’
Variable min max skew c.r. kurtosis c.r.
v26 1.00 7.00 -0.05 -0.33 -1.10 -3.40
v25 1.00 7.00 0.23 1.40 -0.90 -2.80
v24 1.00 7.00 0.12 0.75 -1.10 -3.40
v23 1.00 7.00 -0.07 -0.45 -0.93 -2.88
Multivariate 4.75 5.19
A single-factor measurement model for the construct ‘social capital’ was specified as a
latent variable with four indicators. The model with standardised parameters is illustrated in
Figure 5.16.
202
Figure 5.16. Measurement model for the construct ‘social capital’
The model presented a non-significant chi-square value of .32 (p = .851). The values of
all the baseline comparisons fit indices (NFI, RFI, IFI, TLI, and CFI) were above the cutoff
value of .90 (Bentler, 1990) and the value of RMSEA (< .001) was well below the cutoff
value of .08 (R. Ho, 2006). These indicated a good model fit (see Table 5.39). The
unstandardised regression weight of each indicator corresponded to a significant critical ratio.
All the values of the standardised regression weight (see Figure 5.16) were above the cutoff
value of .50 (Hair et al., 2006) and equal to or above the preferred value of .70 (Hair et al.,
2006). This indicated that all four corresponding items were significant reflective indicators
of the respective construct ‘social capital’. The model converged.
203
Table 5.39
Model fit indices for the construct ‘social capital’
CMIN
Model NPAR CMIN DF P CMIN/DF
Default model 8 .32 2 .851 .16 Saturated model 10 .000 0
Independence model 4 403.38 6 .000 67.23 Baseline comparisons
Model NFI
Delta1 RFI
rho1 IFI
Delta2 TLI
rho2 CFI
Default model 1.00 1.00 1.00 1.01 1.00
Saturated model 1.00 1.00 1.00
Independence model 0.00 0.00 0.00 0.00 0.00
RMSEA
Model RMSEA LO 90 HI 90 PCLOSE
Default model .00 .00 .07 .91 Independence model .54 .49 .58 .00
The proportional factor score weight of each indicator is demonstrated in Table 5.40.
Table 5.40
Proportional factor score weights for the indicators within the construct ‘social capital’
v26 v25 v24 v23 Sum
Factor score weight .15 .24 .18 .34 .91
Proportional factor score weight .16 .26 .20 .37 .99
The scale score for the construct ‘social capital’ then can be calculated by the following
equation:
v26 × .16 + v25 × .26 + v24 × .20 + v23 × .37
Reliability and validity of the construct ‘social capital’ were further checked by four
measures of reliability, namely SMC, construct reliability, variance extracted, and coefficient
H, as well as two measures of validity, namely construct validity and convergent validity.
These measures are reported below.
As demonstrated in Figure 5.16, the SMCs for all the four indicators were above the
cutoff value of .30 (Jöreskog & Sörbom, 1996) and equal to or above the preferred value
204
of .50 (Jöreskog & Sörbom, 1996). Therefore, the construct ‘social capital’ accounted for a
reasonable level of the variance of its indicators. The reliability of all the indicators of the
construct ‘social capital’ represented by the SMC measures was good. The construct
reliability for ‘social capital’ was .61, which was greater than the cutoff value of .50 (Fornell
& Larcker, 1981). The variance of the indicators captured by the construct ‘social capital’ was
greater than that due to the measurement errors. This indicated that the construct reliability
was acceptable. The construct ‘social capital’ accounted for 69.99% of the overall variation in
the four indicators. This was well above the cutoff of 50% (Fornell & Larcker, 1981).
Therefore, the reliability of the indicators of the construct ‘social capital’ represented by the
variance extracted was good. The coefficient H value for the construct ‘social capital’ was .87,
which was well above the cutoff value of .70 (G. R. Hancock & Mueller, 2001). This
represented a high reliability.
The measurement model for the construct ‘social capital’ had a good model fit, which
indicated the unidimensionality of the four indicators. This supported the claim for its
construct validity. As stated earlier, the unstandardised regression weight of all the indicators
of the construct ‘social capital’ corresponded to a significant critical ratio. Therefore, all the
indicators demonstrated a good convergent validity.
Symbolic capital
Originally, five indicators were used to operationalise the construct ‘symbolic capital’. These
indicators were reflective of the degree to which financial wealth, Chinese cultural
dispositions, and Chinese social networks were recognised or valued. This section discusses
whether the five indicators fit into the measurement model for the construct ‘symbolic capital’
and reports on the modification of the measurement model to improve the model fit.
205
The correlation matrix demonstrated that all indicators had a statistically significant
medium correlation with other indicators (p < .001 as shown in Table 5.41), varying within
the cutoff range between .30 and .90 (Field, 2009). This suggested the basis for EFA.
Table 5.41
Correlation matrix for the indicators within the construct ‘symbolic capital’
Correlation v30 v31 v32 v33 v34
v30 1.00 .53 .59 .55 .53
v31 .53 1.00 .72 .66 .63
v32 .59 .72 1.00 .69 .62
v33 .55 .66 .69 1.00 .75
v34 .53 .63 .62 .75 1.00
Sig. v30 v31 v32 v33 v34
v30 .000 .000 .000 .000
v31 .000 .000 .000 .000
v32 .000 .000 .000 .000
v33 .000 .000 .000 .000
v34 .000 .000 .000 .000
As shown in Table 5.42, the KMO measure of sampling adequacy (.86) was well above
the cutoff value of .50 (Kaiser, 1974), which indicated the EFA should yield distinct and
reliable factors. Bartlett’s test of sphericity confirmed the adequacy of magnitude of the
correlations by presenting a statistically significant chi-square value of 662.18 (p < .001).
Table 5.42
KMO and Bartlett's test for the indicators within the construct ‘symbolic capital’
KMO and Bartlett's test
Kaiser-Meyer-Olkin measure of sampling adequacy. .86
Bartlett's test of sphericity Approx. chi-square 662.18
df 10
Sig. .000
An examination of communalities indicated that a reasonable level of variation in the
indicators (between 45% and 73% as highlighted in Table 5.43) was explained by the
construct ‘symbolic capital’. The extraction in all cases explained more variance in an
206
indicator than the initial model (see Table 5.43). This indicated the appropriateness of the
current factor extraction.
Table 5.43
Communalities for the indicators within the construct ‘symbolic capital’
Initial Extraction
v30 .41 .45
v31 .59 .65
v32 .63 .69
v33 .66 .73
v34 .60 .64
Extraction method: Maximum likelihood
There was only one factor with an eigenvalue above 1. Consequently, only one factor
was extracted. This factor accounted for 70.35% of the total variance of the five indicators
(see Table 5.44). The scree plot also supported a one-factor solution (see Figure 5.17).
Table 5.44
Total variance of the construct ‘symbolic capital’ explained by its indicators
Factor Initial eigenvalues Extraction sums of squared loadings
Total % of Variance Cumulative % Total % of Variance Cumulative %
1 3.52 70.35 70.35 3.16 63.25 63.25
2 .53 10.58 80.93
3 .44 8.73 89.65
4 .29 5.71 95.36
5 .23 4.64 100.00
Extraction method: Maximum likelihood
Figure 5.17. Scree plot of the ‘symbolic capital’ scale
207
As highlighted in Table 5.45, all the individual indicators departed significantly from
univariate normal distribution by displaying the critical ratio of kurtosis beyond the cutoff
value of |2| (Field, 2009). Mardia’s coefficient value of 11.30 was much greater than the
cutoff value of 3.00 (Yuan et al., 2002). This indicated significant deviation from multivariate
normality. Therefore, bootstrapping techniques were applied when conducting CFA.
Table 5.45
Assessment of normality for the indicators within the construct ‘symbolic capital’
Variable min max skew c.r. kurtosis c.r.
v34 1.00 7.00 -0.06 -0.40 -0.77 -2.37
v33 1.00 7.00 -0.12 -0.73 -0.91 -2.81
v32 1.00 7.00 0.19 1.15 -1.00 -3.11
v31 1.00 7.00 0.05 0.33 -1.08 -3.36
v30 1.00 7.00 0.31 1.89 -0.95 -2.93
Multivariate 11.30 10.24
A single-factor measurement model for the construct ‘symbolic capital’ was specified
as a latent variable with five indicators. The model with standardised parameters is illustrated
in Figure 5.18.
Figure 5.18. Measurement model for the construct ‘symbolic capital’
208
Model fit indices are demonstrated in Table 5.46. Although all the baseline
comparisons indices were above the cutoff value of .90 (Bentler, 1990), the significant
chi-square value of 23.98 and the RMSEA value of .13, which was above the cutoff value
of .08 (R. Ho, 2006), implied that the model fit could be improved.
Table 5.46
Model fit indices for the construct ‘symbolic capital’
CMIN
Model NPAR CMIN DF P CMIN/DF
Default model 10 23.98 5 .00 4.80 Saturated model 15 .00 0
Independence model 5 669.49 10 .00 66.95 Baseline comparisons
Model NFI
Delta1 RFI
rho1 IFI
Delta2 TLI
rho2 CFI
Default model 0.96 0.93 0.97 0.94 0.97
Saturated model 1.00 1.00 1.00
Independence model 0.00 0.00 0.00 0.00 0.00
RMSEA
Model RMSEA LO 90 HI 90 PCLOSE
Default model 0.13 0.08 0.18 0.01
Independence model 0.54 0.50 0.57 0.00
An examination of the standardised residual covariance matrix (see Table 5.47) and the
modification index (see Table 5.48) indicated that v33 and v34 were of concern. These two
indicators were associated with a high value of standardised residual covariances, a high
value of covariances between indicator error terms, and a high value of regression weights
between indicators (as highlighted in Table 5.47 and 5.48). The online questionnaire items
corresponding to these two indicators were revisited. The item v33 reflected the symbolic
capital produced by the participants’ Chinese heritage while v34 reflected the symbolic
capital produced by the participants’ popularity among their peers. This popularity may or
may not be attributed to participants’ Chinese heritage or CHL. Therefore, v33 was more
meaningful in the current context. The item v34 was tentatively excluded from the model.
209
Table 5.47
Standardised residual covariances for the indicators within the construct ‘symbolic capital’
v34 v33 v32 v31 v30
v34 0.00
v33 0.73 0.00
v32 -0.63 -0.20 0.00
v31 -0.26 -0.34 0.66 0.00
v30 -0.09 -0.28 0.48 -0.07 0.00
Table 5.48
Modification indices for the construct ‘symbolic capital’
Covariances Regression weights
M.I. Par change
e33 <--> e34 12.69 .27
e32 <--> e34 7.54 -.23 e31 <--> e32 8.49 .26
Before testing the fit of the modified model, multivariate normality was checked (see
Table 5.49). Mardia’s coefficient was 6.43, which was greater than the cutoff value of 3.00
(Yuan et al., 2002). This implied the violation of the multivariate normality. Therefore,
bootstrapping was applied when conducting CFA.
Table 5.49
Assessment of normality for the indicators within the construct ‘symbolic capital’ (modified)
Variable min max skew c.r. kurtosis c.r.
v33 1.00 7.00 -0.12 -0.73 -0.91 -2.81
v32 1.00 7.00 0.19 1.15 -1.00 -3.11
v31 1.00 7.00 0.05 0.33 -1.08 -3.36
v30 1.00 7.00 0.31 1.89 -0.95 -2.93
Multivariate 6.43 7.04
A modified measurement model for the construct ‘symbolic capital’ was specified as a
latent variable with four indicators. The model with standardised parameters is illustrated in
Figure 5.19.
M.I. Par change
v33 <--- v34 4.02 .08
210
Figure 5.19. Modified measurement model for the construct ‘symbolic capital’
The model presented a non-significant chi-square value of .82 (p = .665). The values of
all the baseline comparisons fit indices (NFI, RFI, IFI, TLI, and CFI) were above the
cutoff .90 (Bentler, 1990) and the value of RMSEA (< .001) was well below the cutoff .08 (R.
Ho, 2006). These were indicative of a good model fit (see Table 5.50). The unstandardised
regression weight of each indicator corresponded to a significant critical ratio. All the values
of the standardised regression weight (see Figure 5.19) were above the cutoff value of .50
(Hair et al., 2006) and most values of the standardised regression weight were above the
preferred value of .70 (Hair et al., 2006). The only exception was associated with the
standardised regression weight of v30. However, the value of .67 was only marginally
below .70. Therefore, all the four corresponding items were significant reflective indicators of
the respective construct ‘symbolic capital’. The model converged.
211
Table 5.50
Model fit indices for the indicators within the construct ‘symbolic capital’
CMIN
Model NPAR CMIN DF P CMIN/DF
Default model 8 .82 2 .665 .41 Saturated model 10 .00 0
Independence model 4 460.72 6 .00 76.79 Baseline comparisons
Model NFI
Delta1 RFI
rho1 IFI
Delta2 TLI
rho2 CFI
Default model 1.00 1.00 1.00 1.01 1.00
Saturated model 1.00 1.00 1.00
Independence model 0.00 0.00 0.00 0.00 0.00
RMSEA
Model RMSEA LO 90 HI 90 PCLOSE
Default model .00 .00 .10 .79 Independence model .58 .53 .62 .00
Compared to the original model, the modified model had a significantly better fit. The
modified model decreased the chi-square value by 23.16 (23.98-.82) from the original model.
The chi-square difference was a statistically significant one (p < .001) with a degree of
freedom difference of 3 (5-2).
The proportional factor score weight of each indicator is demonstrated in Table 5.51.
Table 5.51
Proportional factor score weights for the indicators within the construct ‘symbolic capital’
v33 v32 v31 v30 Sum
Factor score weight .20 .32 .22 .11 .85
Proportional factor score weight .24 .38 .26 .13 1.01
The scale score for the construct ‘symbolic capital’ then can be calculated by the
following equation:
v33 × .24 + v32 × .38 + v31 × .26 + v30 × .13
Reliability and validity of the construct ‘symbolic capital’ were further checked by four
measures of reliability, namely SMC, construct reliability, variance extracted, and coefficient
H, as well as two measures of validity, namely construct validity and convergent validity.
212
As demonstrated in Figure 5.19, the SMCs for all the four indicators were above the
cutoff value of .30 (Jöreskog & Sörbom, 1996) and most of them were above the preferred
value of .50 (Jöreskog & Sörbom, 1996). The only exception, the value of .45, was
considered not too far away from the cutoff .50. Therefore, the construct ‘symbolic capital’
accounted for a reasonable level of the variance of its indicators. The reliability of all the
indicators of the construct ‘symbolic capital’ represented by the SMC measures was good.
The construct reliability for ‘symbolic capital’ was .63, which was greater than the cutoff
value of .50 (Fornell & Larcker, 1981). The variance of the indicators captured by the
construct ‘symbolic capital’ was greater than that due to the measurement errors. This
indicated that the construct reliability was acceptable. The construct ‘symbolic capital’
accounted for 72.09% of the overall variation in the four indicators. This was well above the
cutoff of 50% (Fornell & Larcker, 1981). Therefore, the reliability of the indicators of the
construct ‘symbolic capital’ represented by the variance extracted was good. The coefficient
H value for the construct ‘symbolic capital’ was .89, which was well above the cutoff value
of .70 (G. R. Hancock & Mueller, 2001). This represented a high reliability.
The modified measurement model for the construct ‘symbolic capital’ had a good
model fit, which indicated the unidimensionality of the four indicators. This supported the
claim for its construct validity. As stated earlier, the unstandardised regression weight of all
the indicators of the construct ‘symbolic capital’ corresponded to a significant critical ratio.
Therefore, all the indicators demonstrated a good convergent validity.
CHL proficiency
Originally, eight indicators were used to measure ‘CHL proficiency’. These indicators were
reflective of four CHL skills, namely listening, speaking, reading, and writing. Each skill was
213
measured by two indicators. This section discusses whether the eight indicators fit into the
measurement model of ‘CHL proficiency’ and reports on the modification of the
measurement model to improve the model fit.
The correlation matrix demonstrated that all the indicators had a statistically significant
medium to strong correlation with other indicators (p < .001 as shown in Table 5.52), varying
within the cutoff range between .30 and .90 (Field, 2009). This suggested the basis for EFA.
Table 5.52
Correlation matrix for the indicators within the construct ‘CHL proficiency’
Correlation v35 v36 v37 v38 39 v40 v41 v42
v35 1.00 .89 .79 .79 .76 .69 .67 .67
v36 .89 1.00 .77 .86 .77 .79 .68 .73
v37 .79 .77 1.00 .88 .82 .72 .75 .69
v38 .79 .86 .88 1.00 .82 .81 .73 .76
v39 .76 .77 .82 .82 1.00 .84 .82 .82
v40 .69 .79 .72 .81 .84 1.00 .74 .85
v41 .67 .68 .75 .73 .82 .74 1.00 .83
v42 .67 .73 .69 .76 .82 .85 .83 1.00
Sig. v35 v36 v37 v38 39 v40 v41 v42
v35 .000 .000 .000 .000 .000 .000 .000
v36 .000 .000 .000 .000 .000 .000 .000
v37 .000 .000 .000 .000 .000 .000 .000
v38 .000 .000 .000 .000 .000 .000 .000
v39 .000 .000 .000 .000 .000 .000 .000
v40 .000 .000 .000 .000 .000 .000 .000
v41 .000 .000 .000 .000 .000 .000 .000
v42 .000 .000 .000 .000 .000 .000 .000
As shown in Table 5.53, the KMO measure of sampling adequacy (.89) was well above
the cutoff value of .50 (Kaiser, 1974), which indicated that the EFA should yield distinct and
reliable factors. Bartlett’s test of sphericity confirmed the adequacy of the magnitude of the
correlations by presenting a statistically significant chi-square value of 2339.37 (p < .001).
214
Table 5.53
KMO and Bartlett's test for the indicators within the construct ‘CHL proficiency’
KMO and Bartlett's test
Kaiser-Meyer-Olkin measure of sampling adequacy. .89
Bartlett's test of sphericity Approx. chi-square 2339.37
df 28
Sig. .000
An examination of communalities indicated that a good level of variation in the
indicators (between 70% and 85% as highlighted in Table 5.54) was explained by the
construct ‘CHL proficiency’. Of concern was that the extraction of all cases explained less
variance in an indicator than the initial model (see Table 5.54). This was an indication that
there might be an issue in the model fit. A closer examination of this issue will be discussed
subsequently.
Table 5.54
Communalities for the indicators within the construct ‘CHL proficiency’
Initial Extraction
v35 .84 .73
v36 .88 .79
v37 .83 .78
v38 .87 .85
v39 .85 .84
v40 .83 .78
v41 .77 .70
v42 .82 .74
Extraction method: Maximum likelihood.
There was only one factor with an eigenvalue above 1, accounting for 80.52% of the
total variance of the eight indicators (see Table 5.55). Thus, only one factor was extracted.
An examination of the scree plot also supported a one-factor solution (see Figure 5.20).
215
Table 5.55
Total variance of the construct ‘CHL proficiency’ explained by its indicators
Factor Initial eigenvalues Extraction sums of squared loadings
Total % of Variance Cumulative % Total % of Variance Cumulative %
1 6.44 80.52 80.52 6.22 77.78 77.78
2 .55 6.87 87.38
3 .32 4.01 91.39
4 .26 3.27 94.66
5 .16 2.02 96.68
6 .11 1.35 98.03
7 .09 1.14 99.17
8 .07 .83 100.00
Extraction method: Maximum likelihood
Figure 5.20. Scree plot of the ‘CHL proficiency’ scale
As highlighted in Table 5.56, all the individual indicators departed significantly from
univariate normal distribution by displaying the critical ratio of skewness, kurtosis, or both
beyond the cutoff value of |2| (Field, 2009). The value of Mardia’s coefficient was 33.52,
which was much higher than the cutoff value of 3.00 (Yuan et al., 2002). This indicated
significant deviation from multivariate normality. Therefore, bootstrapping techniques were
applied when conducting CFA.
216
Table 5.56
Assessment of normality for the indicators within the construct ‘CHL proficiency’
Variable min max skew c.r. kurtosis c.r.
v42 1.00 7.00 0.65 4.01 -0.76 -2.35
v41 1.00 7.00 0.26 1.62 -1.10 -3.40
v40 1.00 7.00 0.58 3.60 -0.94 -2.92
v39 1.00 7.00 -0.03 -0.20 -1.34 -4.16
v38 1.00 7.00 0.12 0.75 -1.30 -4.04
v37 1.00 7.00 -0.14 -0.89 -1.26 -3.91
v36 1.00 7.00 -0.25 -1.52 -1.27 -3.93
v35 1.00 7.00 -0.57 -3.55 -0.91 -2.81
Multivariate 33.52 20.09
A single-factor measurement model for the construct ‘CHL proficiency’ was specified
as a latent variable with eight indicators. The model with standardised parameters is
illustrated in Figure 5.21.
Figure 5.21. Measurement model for the construct ‘CHL proficiency’
The chi-square value of 357.78 corresponded to a significant p (<.001) value. All the
baseline comparisons fit indices were below the cutoff value of .90 (Bentler, 1990). The value
of RMSEA (.27) was well above the cutoff value of .08 (R. Ho, 2006). These were all
indicative of a poor model fit (see Table 5.57). This was not surprising because EFA has
already hinted that there was a model fit problem: The extraction of all cases explained less
variance in an indicator than the initial model.
217
Table 5.57
Model fit indices for the construct ‘CHL proficiency’
CMIN
Model NPAR CMIN DF P CMIN/DF
Default model 16 357.78 20 .000 17.89 Saturated model 36 .00 0
Independence model 8 2375.68 28 .000 84.85 Baseline comparisons
Model NFI
Delta1 RFI
rho1 IFI
Delta2 TLI
rho2 CFI
Default model 0.85 0.79 0.86 0.80 0.86
Saturated model 1.00 1.00 1.00
Independence model 0.00 0.00 0.00 0.00 0.00
RMSEA
Model RMSEA LO 90 HI 90 PCLOSE
Default model 0.27 0.25 0.30 .00 Independence model 0.61 0.59 0.63 .00
An examination of the standardised residual covariance matrix (see Table 5.58) and the
modification index (see Table 5.59) identified some problems. The two highest scores of
standardised residual covariances, covariances between error terms, and regressions weights
were all associated with v35 and v36, and v41 and v42. Moreover, v39 and v40 were of
concern because these two indicators were also associated with high standardised residual
covariances, error terms covariances, and/or regression weights. The questionnaire items
corresponding to these seemingly problematic indicators were revisited. Indicators v35 and
v36 were used to measure CHL listening proficiency; v41 and v42 were used to measure
CHL writing proficiency; and v39 and v40 were used to measure CHL speaking and reading
proficiencies respectively. Each skill was measured by two indicators. A solution to improve
the model fit was then made by deleting one indicator from the measures for each skill. As a
result, CHL proficiency was still measured by participants’ self-reporting proficiency of
listening, speaking, reading, and writing, but each CHL skill was measured by only one
indicator. Indicators v36, v39, v40, and v42 were tentatively removed from the model.
218
Table 5.58
Standardised residual covariances for the indicators within the construct ‘CHL proficiency’
v42 v41 v40 v39 v38 v37 v36 v35
v42 0.00
v41 1.37 0.00
v40 1.05 -0.01 0.00
v39 0.37 0.65 0.38 0.00
v38 -0.38 -0.58 -0.05 -0.27 0.00
v37 -0.87 0.09 -0.75 0.12 0.70 0.00
v36 -0.41 -0.81 0.00 -0.58 0.40 -0.16 0.00
v35 -0.83 -0.49 -0.73 -0.27 0.01 0.39 1.65 0.00
Table 5.59
Modification indices for the construct ‘CHL proficiency’
Covariances Regression weights
M.I. Par change M.I. Par change e41<-->e42 43.57 0.49 v42<---v41 12.13 0.12
e40<-->e42 38.27 0.41 v42<---v40 7.37 0.09
e39<-->e42 7.02 0.16 v42<---v37 5.07 -0.08
e39<-->e41 17.97 0.28 v42<---v35 4.56 -0.08
e39<-->e40 9.15 0.18 v41<---v42 10.31 0.12
e38<-->e42 8.11 -0.17 v40<---v42 9.10 0.10
e38<-->e41 15.69 -0.25 v40<---v37 4.82 -0.07
e38<-->e39 7.65 -0.14 v40<---v35 4.44 -0.08
e37<-->e42 26.02 -0.34 v39<---v41 5.05 0.07
e37<-->e40 24.67 -0.32 v39<---v36 4.24 -0.06
e37<-->e38 34.58 0.33 v38<---v41 4.41 -0.06
e36<-->e42 6.01 -0.17 v38<---v37 6.84 0.08
e36<-->e41 20.02 -0.32 v37<---v42 6.19 -0.09
e36<-->e39 22.60 -0.28 v37<---v40 4.77 -0.07
e36<-->e38 12.06 0.20 v37<---v38 4.26 0.07
e35<-->e42 17.99 -0.28 v36<---v41 5.59 -0.08
e35<-->e41 5.23 -0.16 v36<---v35 23.40 0.18
e35<-->e40 17.46 -0.26 v35<---v42 4.26 -0.07
e35<-->e37 4.80 0.14 v35<---v36 16.99 0.13
e35<-->e36 91.97 0.61
Before testing the fit of the modified model, multivariate normality was checked again
because four indicators had been removed from the model. Mardia’s coefficient was 4.36,
which was greater than the cutoff value of 3.00 (Yuan et al., 2002). This implied the violation
of the multivariate normality (see Table 5.60). Therefore, bootstrapping was applied when
conducting CFA.
219
Table 5.60
Assessment of normality for the indicators within the construct ‘CHL proficiency’ (modified)
Variable min max skew c.r. kurtosis c.r.
v41 1.00 7.00 0.26 1.62 -1.10 -3.40
v38 1.00 7.00 0.12 0.75 -1.30 -4.04
v37 1.00 7.00 -0.14 -0.89 -1.26 -3.91
v35 1.00 7.00 -0.57 -3.55 -0.91 -2.81
Multivariate 4.36 4.77
A modified measurement model for the construct ‘CHL proficiency’ was specified as a
latent variable with four indicators. The model with standardised parameters is illustrated in
Figure 5.22.
Figure 5.22. Modified measurement model for the construct ‘CHL proficiency’
Model fit was improved. The chi-square value of 1.17 corresponded to a
non-significant p value of .557. All the baseline comparisons indices (NFI, RFI, IFI, TLI, and
CFI) were above the cutoff value of .90 (Bentler, 1990). The RMSEA value (< .001) was well
below the cutoff value of .08 (R. Ho, 2006). All of these were indicative of a good model fit
(see Table 5.61). The unstandardised regression weight of each indicator corresponded to a
significant critical ratio. All the values of the standardised regression weight (see Figure 5.22)
were above the cutoff value of .50 (Hair et al., 2006) and the preferred value of .70 (Hair et
220
al., 2006). Therefore, all the four corresponding items were significant reflective indicators of
the respective construct ‘CHL proficiency’. The model converged.
Table 5.61
Model fit indices for the construct ‘CHL proficiency’ (modified)
CMIN
Model NPAR CMIN DF P CMIN/DF
Default model 8 1.17 2 .557 .59 Saturated model 10 .00 0
Independence model 4 786.16 6 .000 131.03 Baseline comparisons
Model NFI
Delta1 RFI
rho1 IFI
Delta2 TLI
rho2 CFI
Default model 1.00 1.00 1.00 1.00 1.00
Saturated model 1.00 1.00 1.00
Independence model 0.00 0.00 0.00 0.00 0.00
RMSEA
Model RMSEA LO 90 HI 90 PCLOSE
Default model 0.00 0.00 0.11 0.71
Independence model 0.75 0.71 0.80 0.00
Compared to the original model, the modified model had a significantly better fit. The
modified model decreased the chi-square value of 356.61 (357.78-1.17) from the original
model. This value was statistically significant (p < .001) with a difference in degrees of
freedom of 18 (20-2).
The proportional factor score weight of each indicator is demonstrated in Table 5.62.
Table 5.62
Proportional factor score weights for the indicators within the construct ‘CHL proficiency’
v41 v38 v37 v35 Sum
Factor score weight .11 .34 .41 .16 1.02
Proportional factor score weight .11 .33 .40 .16 1.00
The scale score for the construct ‘CHL proficiency’ then can be calculated by the
following equation:
v41 × .11 + v38 × .33 + v37 × .40 + v35 × .16
221
Reliability and validity of the construct ‘CHL proficiency’ were further checked by
four measures of reliability, namely SMC, construct reliability, variance extracted, and
coefficient H, as well as two measures of validity, namely construct validity and convergent
validity.
As demonstrated in Figure 5.22, the SMCs for all the four indicators were above the
cutoff value of .30 (Jöreskog & Sörbom, 1996) and the preferred value of .50 (Jöreskog &
Sörbom, 1996). Therefore, the construct ‘CHL proficiency’ accounted for a reasonable level
of the variance of its indicators. The reliability of all the indicators of the construct ‘CHL
proficiency’ represented by the SMC measures was good. The construct reliability for ‘CHL
proficiency’ was .77, which was greater than the cutoff value of .50 (Fornell & Larcker, 1981).
The variance of the indicators captured by the construct ‘CHL proficiency’ was greater than
that due to the measurement errors. This indicated that the construct reliability was good. The
construct ‘CHL proficiency’ accounted for 82.56% of the overall variation in the four
indicators. This was well above the cutoff of 50% (Fornell & Larcker, 1981). Therefore, the
reliability of the indicators of the construct ‘CHL proficiency’ represented by the variance
extracted was good. The coefficient H value for the construct ‘CHL proficiency’ was .95,
which was well above the cutoff value of .70 (G. R. Hancock & Mueller, 2001). This
represented a high reliability.
The modified measurement model for the construct ‘CHL proficiency’ had a good
model fit, which indicated the unidimensionality of the four indicators. This supported the
claim for its construct validity. As stated earlier, the unstandardised regression weight of all
the indicators of the construct ‘CHL proficiency’ corresponded to a significant critical ratio.
Therefore, all the indicators demonstrated a good convergent validity.
222
Section summary
In this section, the six theoretical constructs, namely ‘Chineseness’, ‘economic capital’,
‘cultural capital’, ‘social capital’, ‘symbolic capital’, and ‘CHL proficiency’, were specified
as latent variables with their corresponding indicators. The initial EFA and the subsequent
CFA improved the model fit to a reasonably good level. These constructs and their respective
indicators included in the (modified) measurement models are summarised in Table 5.63.
Table 5.63
Item sets and their corresponding constructs in the (modified) measurement models
Construct Indicators Item description
Chineseness
(modified)
v2 Handwriting tells you a lot about a person’s character. To what extent do you
agree with this statement?
v3 Children must be taught to paint, dance, or play musical instruments. To what
extent do you agree with this statement?
v4 Academic education is the most important thing in school. To what extent do
you agree with this statement?
v7 People who are senior in age and/or position should be addressed by their
title plus surname rather than their first name. To what extent do you agree
with this statement?
v9 Confucianism values the virtue of frugality. Confucius said that frugality is a
common character of people with virtue. He also said that frugality is very
important in family life. In general, how much degree would you agree with
these values?
Economic
capital
v11 Money is not an issue if I wish to travel to China. To what extent do you
agree with this statement?
v12 I can afford to study in Chinese language schools as many hours as I wish. To
what extent do you agree with this statement?
v13 I can afford to employ private tutors to teach me Chinese if I wish. To what
extent do you agree with this statement?
v14 Money is never an issue if I want to buy Chinese language learning materials,
such as textbooks, dictionaries, and tapes. To what extent do you agree with
this statement?
Cultural
capital
(modified)
v17 I always keep up to date with current Chinese affairs by watching TV,
listening to the radio, reading newspapers, or surfing online. To what extent
do you agree with this statement?
v18 I always go to venues, such as libraries, galleries, museums, theatres, or
concerts, if they feature Chinese culture. To what extent do you agree with
this statement?
v19 I read a lot of books about China. To what extent do you agree with this
statement?
v20 I have invested a lot of time in practising Chinese cultural activities, such as
learning musical instruments, calligraphy, or painting. To what extent do you
agree with this statement?
Social
capital
v23 Most of my friends are of Chinese descent. To what extent do you agree with
this statement?
223
v24 I am a member of social groups, such as a church or club, which mostly
include members of Chinese descent. To what extent do you agree with this
statement?
v25 I tend to mix exclusively with Chinese social groups. To what extent do you
agree with this statement?
v26 I think it is necessary to learn Chinese language in order to socialise with my
Chinese peers. To what extent do you agree with this statement?
Symbolic
capital
(modified)
v30 The wealth of my family is well-known in Chinese communities. To what
extent do you agree with this statement?
v31 My Chinese language competency is well regarded. To what extent do you
agree with this statement?
v32 Learning Chinese has increased my status in Chinese communities. To what
extent do you agree with this statement?
v33 People value my Chinese heritage. To what extent do you agree with this
statement?
CHL
proficiency
(modified)
v35 I can easily understand my family members and friends when they talk to me
in Chinese. To what extent do you agree with this statement?
v37 I can easily handle complex situations in Chinese, such as banking, arguing,
purchasing a house or a car. To what extent do you agree with this statement?
v38 I can read Chinese language textbooks easily. To what extent do you agree
with this statement?
v41 I can always write Chinese characters and Chinese words correctly. To what
extent do you agree with this statement?
5.2.2 Hypothesis testing: SEM
In the previous section (5.2.1), modified measurement models for the theoretical constructs
and their respective indicators were specified by CFA. As such, CFA provides the foundation
for hypothesis testing. In this section, the primary investigation will shift to the examination
of the relationships between the theoretical constructs. Drawing insights from the theoretical
framework, the structural models can specify the nature and magnitude of the relationships
between constructs. Therefore, the structural models function to test the hypotheses, and
further, test whether the structural models accord with the theoretical framework. The testing
of Hypotheses One and Two follows.
224
H01. The habitus of ‘Chineseness’ of young Chinese Australian adults has no
statistically significant impact on their CHL proficiency.
The relationship between ‘Chineseness’ and ‘CHL proficiency’ was specified in the structural
model, with the standardised parameters demonstrated in Figure 5.23.
Figure 5.23. Structural model for the relationship between ‘Chineseness’ and ‘CHL proficiency’
The model had a significant chi-square value of 57.92. However, the ratio between the
chi-square value and the degrees of freedom (57.92/26 = 2.23) was not very high; all the
baseline comparison fit indices were above the cutoff value of .90 (Bentler, 1990); and the
RMSEA value of .07 was below the cutoff value of .08 (R. Ho, 2006). It can be argued that
the model had a reasonably good fit given its complexity (see Table 5.64). The
unstandardised regression weight of each indicator corresponded to a significant critical ratio.
All the values of the standardised regression weight were above the cutoff value of .50 (Hair
et al., 2006) and equal to or above the preferred value of .70 (Hair et al., 2006). The model
converged.
225
Table 5.64
Model fit indices for the structural model: ‘Chineseness’ and ‘CHL proficiency’
CMIN
Model NPAR CMIN DF P CMIN/DF
Default model 19 57.92 26 .000 2.23
Saturated model 45 .00 0
Independence model 9 1446.81 36 .000 40.19
Baseline comparisons
Model NFI
Delta1
RFI
rho1
IFI
Delta2
TLI
rho2 CFI
Default model .96 .95 .98 .97 .98
Saturated model 1.00 1.00 1.00
Independence model .00 .00 .00 .00 .00
RMSEA
Model RMSEA LO 90 HI 90 PCLOSE
Default model .07 .05 .10 .06
Independence model .41 .40 .43 .00
The results indicated that the construct ‘Chineseness’ was significantly and positively
related to the construct ‘CHL proficiency’ (r = .71, see Figure 5.23). The null hypothesis H01
can be rejected. ‘Chineseness’ of young Chinese Australian adults had a statistically
significant strong positive impact on their CHL proficiency. When young Chinese Australian
adults had a stronger sense of ‘Chineseness’, they tended to have a higher level of CHL
proficiency. About 50% of the variance of ‘CHL proficiency’ was accounted for by the
variance of ‘Chineseness’ (see Figure 5.23). The remaining 50% could not be explained by
the model and was thus attributed to other factors.
In line with Bourdieu’s capital metaphor, Garnham and Williams (1993) argue that the
investment of all forms of capital leads to a given agent’s choice of cultivating a language. As
such, young Chinese Australian adults’ investment of various forms of capital may explain
part of the variance in their CHL proficiency. Accordingly, a null hypothesis follows.
226
H02. The investment of capital of young Chinese Australian adults has no statistically
significant impact on their CHL proficiency.
According to Bourdieu (1986, 1991), capital has four forms, namely economic capital,
cultural capital, social capital, and symbolic capital. The theoretical construct ‘capital’ was
then treated as a latent variable with four indicators, with each indicator representing one
form of capital. Each capital form was treated as an ‘item parcel’. This strategy of ‘item
parcelling’ combines indicators into sets of several composite variables by either summing or
averaging the indicator score of the combined indicators (Hair et al., 2006). As illustrated in
the earlier sections, the scale score for the four composite variables, namely ‘economic
capital’, ‘cultural capital’, ‘social capital’, and ‘symbolic capital’, can be computed as
continuous variables by multiplying the individual’s raw score on each indicator by the
proportionally weighted factor score of each indicator and summing. The formulae for
computation are summarised in Table 5.65.
Table 5.65
Formulae for computation of the four composite variables
Economic capital v14 × .21 + v13 × .31 + v12 × .36 + v11 × .12
Cultural capital v20 × .18 + v19 × .29 + v18 × .35 + v17 × .19
Social capital v26 × .16 + v25 × .26 + v24 × .20 + v23 × .37
Symbolic capital v33 × .24 + v32 × .38 + v31 × .26 + v30 × .13
There are several methodological considerations regarding the rationale for item
parcelling. First of all, indicator number was of concern. The theoretical construct ‘capital’
was reflected by 16 indicators with each capital form reflected by four reliable and valid
indicators. Applications involving more than 15 reflective indicators can call for item
parcelling (Hair et al., 2006). Secondly, the 16 indicators displayed high internal consistency
reliability with a Cronbach’s α value of .95. Cronbach’s α above the cutoff value of .90
suggested the basis for item parcelling (Hair et al., 2006). Thirdly, the parcels, or the
227
composite variables contained groups of indicators with the most conceptual similarity. This
is true because each parcel represented a form of capital with its four reflective indicators.
The previous reliability test, EFA, and CFA have confirmed that the reflective indicators
within their corresponding parcels shared a conceptual basis. Moreover, the four parcels
suggested unidimensionality because EFA run on these four parcels demonstrated that they all
loaded on one factor. This indicated that the four capital forms were represented by one
construct, that is, ‘capital’. The EFA results are discussed as follows.
The correlation matrix demonstrated that all capital forms had a statistically significant
medium correlation with other capital forms (p < .001 as shown in Table 5.66), varying
within the cutoff range between .30 and .90 (Field, 2009). This suggested the basis for
conducting EFA.
Table 5.66
Correlation matrix for the indicators within the construct ‘capital’
Correlation Economic Cultural Social Symbolic
Economic 1.00 .64 .57 .65
Cultural .64 1.00 .72 .78
Social .57 .72 1.00 .74
Symbolic .65 .78 .74 1.00
Sig. Economic Cultural Social Symbolic
Economic .000 .000 .000
Cultural .000 .000 .000
Social .000 .000 .000
Symbolic .000 .000 .000
As shown in Table 5.67, the KMO measure of sampling adequacy (.84) was well above
the cutoff value of .50 (Kaiser, 1974), which indicated that the EFA should yield distinct and
reliable factors. Bartlett’s test of sphericity confirmed the adequacy of the magnitude of the
correlations by presenting a statistically significant chi-square value of 564.87 (p < .001).
228
Table 5.67
KMO and Bartlett's test for the indicators within the construct ‘capital’
KMO and Bartlett's test
Kaiser-Meyer-Olkin measure of sampling adequacy. .84
Bartlett's test of sphericity Approx. chi-square 564.87
df 6
Sig. .000
As highlighted in Table 5.68, a good level of variation in the four forms of capital
(between 51% and 81%) was explained by the construct ‘capital’. The extraction of all cases
explained greater variance in a capital form than the initial model.
Table 5.68
Communalities for the indicators within the construct ‘capital’
Initial Extraction
Economic .47 .51
Cultural .67 .76
Social .61 .68
Symbolic .70 .81
Extraction method: Maximum likelihood.
There was only one factor with an eigenvalue above 1, accounting for 76.32% of the
total variance of the four forms of capital (see Table 5.69). As a result, only one factor was
extracted. The scree plot also supported a one-factor solution (see Figure 5.24).
Table 5.69
Total variance of the construct ‘capital’ explained by its indicators
Factor Initial eigenvalues Extraction sums of squared loadings
Total % of Variance Cumulative % Total % of Variance Cumulative %
1 3.05 76.32 76.32 2.76 68.96 68.96
2 .45 11.35 87.67
3 .28 6.96 94.63
4 .21 5.37 100.00
Extraction method: Maximum likelihood
229
Figure 5.24. Scree plot of the ‘capital’ scale
A measurement model for the construct ‘capital’ was evaluated by CFA. As shown in
Table 5.70, the data violated the multivariate normal distribution assumption because
Mardia’s coefficient was 4.75, which was greater than the cutoff value of 3.00 (Yuan et al.,
2002). Bootstrapping techniques were applied when conducting CFA.
Table 5.70
Assessment of normality for the indicators within the construct ‘capital’
Variable min max skew c.r. kurtosis c.r.
Symbolic capital 1.00 7.00 0.14 0.86 -1.01 -3.11
Social capital 1.00 7.00 -0.01 -0.05 -0.98 -3.05
Cultural capital 1.00 7.00 0.22 1.35 -0.97 -2.99
Economic capital 1.00 7.00 0.01 0.08 -1.01 -3.11
Multivariate 4.75 5.19
Item parcelling has the potential to improve model fit simply because it reduces the
complexity of the model, and models with fewer indicators have the potential for better fit
(Landis, Beal, & Tesluk, 2000). This is particularly true for the item parcelling in the current
study. As shown in Table 5.71, the model achieved a good fit. Chi-square value (.97)
corresponded to a non-significance level (p = .616). All the baseline comparisons model fit
indices were above the cutoff value of .90 (Bentler, 1990). RMSEA value (< .001) was well
below the cutoff value of .08 (R. Ho, 2006).
230
Table 5.71
Model fit indices for the construct ‘capital’
CMIN
Model NPAR CMIN DF P CMIN/DF
Default model 8 .97 2 .616 .49
Saturated model 10 .000 0
Independence model 4 570.27 6 .000 95.04
Baseline comparisons
Model NFI
Delta1
RFI
rho1
IFI
Delta2
TLI
rho2 CFI
Default model 1.00 1.00 1.00 1.01 1.00
Saturated model 1.00 1.00 1.00
Independence model .000 .000 .000 .000 .000
RMSEA
Model RMSEA LO 90 HI 90 PCLOSE
Default model .00 .00 .11 .75
Independence model .64 .60 .69 .00
The standardised parameters are demonstrated in Figure 5.25. The unstandardised
regression weight of each indicator corresponded to a significant critical ratio. All the values
of the standardised regression weight were above the cutoff value of .50 (Hair et al., 2006)
and the preferred value of .70 (Hair et al., 2006). The model converged.
Figure 5.25. Measurement model for the construct ‘capital’
231
Table 5.72 demonstrates the proportional factor score weight of each form of capital.
Table 5.72
Proportional factor score weights for the indicators within the construct ‘capital’
Economi
c capital
Cultural
capital
Social
capital
Symbolic
capital Sum
Factor score weight .11 .28 .21 .37 .97
Proportional factor score weight .11 .29 .22 .38 1.00
The scale score for the construct ‘capital’ can then be computed by the following
equation:
economic capital × .11 + cultural capital × .29 + social capital × .22 + symbolic capital × .38
Four measures of reliability, namely SMC, construct reliability, variance extracted, and
coefficient H, as well as two measures of validity, namely construct validity and convergent
validity, are reported below.
As demonstrated in Figure 5.25, the SMCs for all the four indicators were above the
cutoff value of .30 (Jöreskog & Sörbom, 1996) and the preferred value of .50 (Jöreskog &
Sörbom, 1996). Therefore, the construct ‘capital’ accounted for a reasonable level of the
variance of its indicators, namely four forms of capital. The reliability of all the indicators of
the construct ‘capital’ represented by the SMC measures was good. The construct reliability
for ‘capital’ was .69, which was greater than the cutoff value of .50 (Fornell & Larcker, 1981).
The variance of four forms of capital captured by the construct ‘capital’ was greater than that
due to the measurement errors. This indicated that the construct reliability was acceptable.
The construct ‘capital’ accounted for 76.32% of the overall variation in four forms of capital.
This was well above the cutoff of 50% (Fornell & Larcker, 1981). Therefore, the reliability of
the indicators of the construct ‘capital’ represented by the variance extracted was good. The
232
coefficient H value for the construct ‘capital’ was .91, which was well above the cutoff value
of .70 (G. R. Hancock & Mueller, 2001). This represented a high reliability.
The measurement model for the construct ‘capital’ had a good model fit, which
indicated the unidimensionality of the four indicators. This supported the claim for its
construct validity. As stated earlier, the unstandardised regression weight of all the indicators
of the construct ‘capital’ corresponded to a significant critical ratio. Therefore, all the
indicators demonstrated a good convergent validity.
The relationship between ‘capital’ and ‘CHL proficiency’ was then specified in the
structural model, with the standardised parameters demonstrated in Figure 5.26.
Figure 5.26. Structural model for the relationship between ‘capital’ and ‘CHL proficiency’
Although the model had a significant chi-square value of 66.30, the ratio of chi-square
value and degree of freedom (3.50) was not very high. Although the RMSEA value of .10
was above the cutoff value of .08 (R. Ho, 2006), all the baseline comparison fit indices were
above the cutoff value of .90 (Bentler, 1990). Particularly, CFI and TLI, the two most widely
referenced indices (Hair et al., 2006), had relatively high values of .97 and .96 respectively.
233
Therefore, it can be argued that the model had a reasonably good fit given its complexity (see
Table 5.73). The unstandardised regression weight of each indicator corresponded to a
significant critical ratio. All the values of the standardised regression weight (see Figure 5.26)
were above the cutoff value of .50 (Hair et al., 2006) and most of the standardised regression
weights were above the preferred value of .70 (Hair et al., 2006). The only exception was the
standardised regression weight associated with ‘economic capital’. However, the value of .69
was only marginally lower than the preferred value of .70. Therefore, the model was
considered to be converged.
Table 5.73
Model fit indices for the structural model: ‘Capital’ and ‘CHL proficiency’
CMIN
Model NPAR CMIN DF P CMIN/DF
Default model 17 66.30 19 .000 3.50
Saturated model 36 .000 0
Independence model 8 1648.92 28 .000 58.89
Baseline comparisons
Model NFI
Delta1
RFI
rho1
IFI
Delta2
TLI
rho2 CFI
Default model .96 .94 .97 .96 .97
Saturated model 1.00 1.00 1.00
Independence model .00 .00 .00 .000 .00
RMSEA
Model RMSEA LO 90 HI 90 PCLOSE
Default model .10 .08 .13 .001
Independence model .50 .48 .52 .000
The results indicated that the construct ‘capital’ was significantly and positively related
to the construct ‘CHL proficiency’ (r = .86, see Figure 5.26). The null hypothesis H02 can be
rejected. The investment of capital of young Chinese Australian adults had a statistically
significant strong positive impact on their CHL proficiency. When young Chinese Australian
adults put more investment of different forms of capital into CHL learning, they tended to
achieve a higher level of CHL proficiency. ‘Capital’ can therefore explain part of the variance
of ‘CHL proficiency’ that could not be explained by ‘Chineseness’.
234
What is also of interest is that the variance of ‘capital’ was reflected in the variance of
different forms of capital at different levels. Compared to other capital forms, symbolic
capital had the most variance caused by the variance of capital (.93, see Figure 5.26). This is
because symbolic capital is perhaps the most valuable form of accumulation in a society
(Bourdieu, 1977b). It will be recalled that economic, cultural, and social capital all had
stronger correlations with symbolic capital than their correlations with other forms of capital
(as shown in Table 5.66). This is because participants’ investment of economic, cultural, and
social capital, when perceived and recognised as legitimate (Bourdieu, 1985b, 1989b, 1991,
2000), can all exist and act through the form of symbolic capital.
According to Bourdieu (1991), people make choices about different aspects of a
particular language according to the amount of different resources (capital) as well as the
dispositions (habitus) that they have in a given field. The theoretical framework developed in
Chapter Three extended this Bourdieusian stance to people’s choices about different
languages according to their habitus in a field and their capital valued by that field. The data
confirmed this theoretical framework and justified the understandings that both participants’
investment of ‘capital’ and habitus of ‘Chineseness’ contributed to their ‘CHL proficiency’.
5.2.3 Habitus and capital working together: Regression
In the previous section (5.2.2), it was concluded that participants’ ‘Chineseness’ (habitus) and
their various resources (capital) made statistically significant positive contributions to their
CHL proficiency, when the relationship between ‘Chineseness’ and ‘CHL proficiency’ and
the relationship between ‘capital’ and ‘CHL proficiency’ were examined separately. This
section extended the scope and depth of the findings from the quantitative data analysis by
investigating the co-contribution of ‘habitus’ and ‘capital’ to CHL proficiency. Multiple
235
regression was used to model the relationship. ‘CHL proficiency’ was treated as the
dependent variable and ‘Chineseness’ and ‘capital’ were treated as independent variables. A
dependent variable refers to the variable whose variance is the presumed effect of, or
response to, the variance of the variables that have been manipulated; while an independent
variable, or a predictor variable, refers to the manipulated variable that is the presumed cause
of the dependent variable (Field, 2009; Hair et al., 2006).
As illustrated in the earlier sections, the scale score for ‘CHL proficiency’, ‘capital’,
and ‘Chineseness’ can be computed as continuous variables by multiplying the individual’s
raw score on each indicator by the proportionally weighted factor score of each indicator and
summing. The formulae for computation are summarised in Table 5.74.
Table 5.74
Formulae for computation of the constructs ‘Chineseness’, ‘capital’, and ‘CHL proficiency’
Chineseness v9 × .18 + v7 × .21 + v4 × .21 + v3 × .16 + v2 × .23
Capital economic capital × .11 + cultural capital × .29 + social capital × .22 +
symbolic capital × .38
CHL proficiency v41 × .11 + v38 × .33 + v37 × .40 + v35 × .16
As shown in Table 5.75, any pair of variables had a significantly medium positive
correlation. Multicollinearity was not suggested at this stage because there was no correlation
above .90 (Field, 2009).
Table 5.75
Correlation matrix for the variables
CHL proficiency Chineseness Capital
Pearson Correlation
CHL proficiency 1.00 .64 .79
Chineseness .64 1.00 .76
capital .79 .76 1.00
Sig. (1-tailed)
CHL proficiency . .000 .000
Chineseness .000 . .000
capital .000 .000 .
236
The way in which independent variables are entered into the model can have a great
impact on the values of the regression coefficients. There are three methods for entering
independent variables into a model. In the hierarchical blockwise entry method, independent
variables should be entered into the model in order of their importance in predicting the
dependent variable. According to Bourdieu, habitus and capital contribute to certain language
choices in a given field, with neither habitus nor capital primary or dominant. Therefore, the
hierarchical blockwise entry was considered inappropriate for the current study. In the
stepwise entry method, the order in which independent variables are entered into the model is
based on a purely mathematical criterion. This takes many important methodological
decisions out of the hands of the researcher. Therefore, stepwise entry method is best avoided
except for exploratory model building (Field, 2009). Unlike the first two methods, forced
entry is a method in which independent variables are forced into the model simultaneously.
This method is believed to be the only appropriate method for theory testing (Field, 2009;
Studenmund & Cassidy, 1987). As such, entering ‘Chineseness’ and ‘capital’ at the same time
into the model can test whether the data fit Bourdieu’s theory. Therefore, the forced entry
method was selected.
As shown in Table 5.76, 62% of the variation of ‘CHL proficiency’ was explained by
the variation of ‘Chineseness’ and ‘capital’. For any two observations, the residual terms
should be uncorrelated or independent (Field, 2009). This assumption can be tested by
Durbin-Watson test. The test statistics can vary between 0 and 4 with a value of 2 meaning
that the residuals are uncorrelated (Durbin & Watson, 1950, 1951). As a rule of thumb, values
less than 1 or greater than 3 should raise concern (Field, 2009). The Durbin-Watson statistics
was 1.97, which was very close to 2 (see Table 5.76). This indicated that the assumption was
adequately met.
237
Table 5.76
Model summary
Model R R
square
Adjusted R
square
Std. Error of
the estimate
R square
change F change df1 df2
Sig. F
change
Durbin-
Watson
1 .79a .62 .62 1.14 .62 187.67 2 227 .000 1.97
a. Predictors: (Constant), capital, Chineseness
b. Dependent variable: CHL proficiency
ANOVA results were demonstrated in Table 5.77. It indicated that the model was a
significant fit of the data overall. ‘Chineseness’ and ‘capital’ together made a statistically
significant contribution to ‘CHL proficiency’ (F = 187.67, p < .001).
Table 5.77
ANOVA results
Model Sum of squares df Mean square F Sig.
1 Regression 485.77 2 242.87 187.67 .000a
Residual 293.79 227 1.29
Total 779.56 229
a. Predictors: (Constant), capital, Chineseness b. Dependent variable: CHL proficiency
As shown in Table 5.78, when ‘Chineseness’ and ‘capital’ worked together, ‘capital’
was a significant contributor to ‘CHL proficiency’ when ‘Chineseness’ was held constant (t =
11.44, p < .001); while ‘Chineseness’ was not a significant contributor to ‘CHL proficiency’
when ‘capital’ was held constant (t = 1.40, p = .163). This issue was examined from a
statistical perspective.
Table 5.78
Model parameters
Model
Unstandardised
coefficients
Standardised
coefficients t Sig.
Collinearity
statistics
B Std. Error Beta Tolerance VIF
1 (Constant) .24 .24 1.02 .310
Chineseness .12 .08 .09 1.40 .163 .42 2.39
capital .92 .08 .72 11.44 .000 .42 2.39
Dependent variable: CHL proficiency
238
The Variance Inflation Factor (VIF) and the tolerance statistics can be used as measures
for multicollinearity. The largest VIF greater than 10, the average VIF substantially greater
than 1, and/or the tolerance less than .20 are indicative of multicollinearity (Field, 2009). For
the current model, VIF values of 2.39 were well below the cutoff of 10 and the tolerance
statistics of .42 were above .20 (see Table 5.78). Multicollinearity was not diagnosed at this
stage.
However, a closer examination of the collinearity diagnostics identified a problem. The
variance proportions for each independent variable should be distributed across different
dimensions. For the current model, ‘Chineseness’ and ‘capital’ had most of their variance
loading onto the same dimension, with 95% of the variance of ‘Chineseness’ and 79% of the
variance of ‘capital’ on dimension 3 (see Table 5.79). Multicollinearity between ‘Chineseness’
and ‘capital’ was diagnosed. The correlation matrix was revisited (see Table 5.75). The
correlation between ‘Chineseness’ and ‘capital’ (r = .76, p < .001) was stronger than the
correlation between ‘Chineseness’ and ‘CHL proficiency’ (r = .64, p < .001). Since the
correlation between the two independent variables (‘Chineseness’ and capital) was stronger
than that between the independent variable (‘Chineseness’) and the dependent variable (CHL
proficiency), multicollinearity between ‘Chineseness’ and ‘capital’ was diagnosed from a
statistical perspective. This will be discussed from a Bourdieusian theoretical perspective in
Chapter Seven.
Table 5.79
Collinearity diagnostics
Dimension Eigenvalue Condition index Variance proportions
(Constant) Chineseness Capital
1 2.90 1.00 .01 .01 .01
2 .07 6.41 .92 .05 .20
3 .03 10.46 .07 .95 .79
Dependent variable: CHL proficiency
239
The multicollinearity and independence of residuals have been checked earlier. At the
final stage in this regression analysis, the assumptions of normality and homoscedasticity (the
residuals at each level of the independent variable have similar variances) were checked. The
distribution of the standardised residuals of ‘CHL proficiency’ was roughly normal with a
nearly bell-shaped curve, although there was a slight left-skewness. The normal probability
plot also demonstrated a roughly normal distribution. Although the points of the observed
residuals distributed up and down along the normal distribution line, there was no substantial
deviation from the straight line (see Figure 5.27).
Figure 5.27. Histogram and P-P plots of the standardised residuals of the ‘CHL proficiency’
The scatter plot of the standardised residuals against the standardised predicted values
as well as that of the studentised residuals against the standardised predicted values
demonstrated a random array of dots evenly dispersed around zero. This indicated that the
assumptions of normality and linearity were met (see Figure 5.28).
240
Figure 5.28. Scatter plots of the standardised and the studentised residuals of ‘CHL proficiency’
against the standardised predicted values of ‘CHL proficiency’
The partial plots were the scatter plots of the residuals of ‘CHL proficiency’ and those
of ‘capital’ and ‘Chineseness’, when ‘CHL proficiency’ was regressed separately on ‘capital’
and ‘Chineseness’ (see Figure 5.29). The plots demonstrated a strong positive relationship
between ‘CHL proficiency’ and ‘capital’ and between ‘CHL proficiency’ and ‘Chineseness’.
There were no obvious outliers on the plots. Although the relationship between ‘CHL
proficiency’ and ‘Chineseness’ looked less linear than that between ‘CHL proficiency’ and
‘capital’, the clouds of dots were evenly spaced around the gradient line in both graphs.
These were indicative of homoscedasticity.
Figure 5.29. Partial plots of the residuals of ‘CHL proficiency’ against ‘capital’ and ‘Chineseness’
241
Despite the glitch of multicollinearity between ‘capital’ and ‘Chineseness’, the
assumptions of independence of residuals, normality, and homoscedasticity have been met. It
can probably be assumed that the model appeared in most senses to be both accurate for the
sample and generalisable to the population.
5.3 Chapter summary
This chapter reported the strategies of the quantitative data analysis in detail. At the first stage
of the pilot study, face validity was improved by adjusting the wording of some items
according to the feedback from five participants. The second stage of the pilot study worked
with the data produced by 38 complete cases. The inter-item reliability of each theoretical
construct was checked in turn. Taking account of both statistical and theoretical
considerations, problematic items were excluded from the model. The overall inter-item
reliability was considered to be improved.
In the main study, data were produced by 230 complete cases. EFA was first applied to
each theoretical construct. Results from EFA suggested the basis for CFA. Each theoretical
construct was then specified in a measurement model by CFA as a latent variable with its
respective indicators. Both statistical and theoretical considerations were taken into account
when excluding problematic indicators from the model to improve the model fit. Under the
modified measurement model, an equation to calculate the scale score of each theoretical
construct was offered. Four measures of reliability, namely SMC, construct reliability,
variance extracted, and coefficient H, as well as two measures of validity, namely construct
validity and convergent validity, were also checked to confirm the reliability and validity of
the modified measurement models.
242
Relationships between these theoretical constructs were specified in the structural
models according to the first part of the theoretical framework (Figure 4.1). The structural
models served the hypothesis testing. The results from SEM rejected both null hypotheses
and led to the conclusion that both ‘Chineseness’ (r = .71, p < .001) and ‘capital’ (r = .86, p
< .001) had statistically significant positive contributions to ‘CHL proficiency’. The analyses
demonstrated that ‘Chineseness’ (habitus) of young Chinese Australian adults and various
resources (capital) available to them explained their CHL proficiency. This was interpreted
through Bourdieu’s argument that language choice can be attributed to agents’ dispositions
(habitus) and various resources (capital) available to them.
In the multiple regression analysis, ‘CHL proficiency’ was regressed on ‘Chineseness’
and ‘capital’. ‘Chineseness’ and ‘capital’ accounted for 62% of the variance of ‘CHL
proficiency’. Statistically, there was a problem of multicollinearity between ‘Chineseness’
and ‘capital’. Theoretically, this multicollinearity aligns with Bourdieu’s (1973; Bourdieu,
1985a) argument that habitus can be understood as a form of capital when valued and
recognised in a given field. The connection between participants’ ‘Chineseness’ and cultural
capital may explain the fact that the contribution made by ‘Chineseness’ was actually
embedded in the contribution made by ‘capital’. This will be further discussed in the final
chapter.
This chapter has reported on the investigation of the first research question. If
‘Chineseness’ qua habitus and accessible resources as ‘capital’ can produce CHL proficiency,
it will be meaningful to go one step further and ask: Does investment in CHL have any
returns? What are the returns on this investment? This leads to the second research question:
How do young Chinese Australian adults understand their CHL learning in relation to
(potential) profits produced by this linguistic capital in given fields? This will be explored
through the qualitative phase reported in the next chapter.
243
Chapter 6: Qualitative Data Analysis and Findings
This chapter presents the analysis of the interview data and reports the findings from the
qualitative phase of the research. The focus of this chapter addresses RQ2: How do young
Chinese Australian adults understand their CHL learning in relation to (potential) profits
produced by this linguistic capital in given fields? The examination of this research question
occurs within the second part of the theoretical framework summarised in Figure 4.4 and
explores the meanings that participants attached to CHL in terms of its profit in capital and
habitus (re)production within particular fields. This chapter is organised in five sections.
Section 6.1 explains the rationale behind the purposive selection of interview participants.
Section 6.2 draws portraits of the five participants and outlines their key characteristics
relevant to this study. Section 6.3 details the interview data analysis and presents the findings.
Section 6.4 considers how the findings from the qualitative phase talk back to the findings
from the initial quantitative phase. Section 6.5 concludes this chapter with a brief summary.
6.1 How were the five participants selected?
Nineteen participants indicated at the end of the online questionnaire that they were willing to
participate in an interview and left their contact details for arranging a time and place of the
interview. Their respondent numbers in the online questionnaire were listed as follows: 30, 46,
49, 59, 66, 77, 79, 87, 89, 91, 92, 120, 129, 134, 139, 141, 225, 227, and 228. These
respondents became candidates for the interview.
As stated in the description of the design of the interview in Chapter Four, Patton’s
(1990) extreme sampling was applied at the first stage to illuminate dimensions behind the
244
average cases. Outliers in the initial quantitative phase were considered as extreme cases for
the purposes of the qualitative phase. To identify these outliers, ‘CHL proficiency’ was
regressed on ‘Chineseness’ and ‘capital’ respectively. Outliers outside ±2 standard deviations
were highlighted. This is because only about 5% of the cases will fall outside ±2 standard
deviations in a normal distribution. Table 6.1 demonstrated the outliers in the regression
relationship between ‘CHL proficiency’ and ‘Chineseness’ as well as in that between ‘CHL
proficiency’ and ‘capital’.
Table 6.1
Case-wise diagnostics
Chineseness-CHL proficiency
Respondent no. Std. residual CHL proficiency Predicted value Residual
5 2.067 7.00 4.0542 2.94583
9 -2.314 1.27 4.5683 -3.29831
29 2.238 6.78 3.5898 3.19021
33 2.100 5.72 2.7274 2.99263
67 2.337 6.49 3.1586 3.33142
83 -2.800 1.00 4.9912 -3.99122
88 -2.826 1.71 5.7375 -4.02755
114 -2.197 2.15 5.2815 -3.13146
122 -2.599 1.32 5.0244 -3.70439
130 -2.269 1.16 4.3942 -3.23416
134 -2.174 1.32 4.4190 -3.09904
142 2.055 7.00 4.0708 2.92925
218 -2.406 2.15 5.5800 -3.42999
228 2.299 7.00 3.7225 3.27753
Capital-CHL proficiency
Respondent no. Std. residual CHL proficiency Predicted value Residual
5 2.82 7.00 3.78 3.22
27 2.42 6.27 3.51 2.76
42 2.94 6.78 3.43 3.35
44 2.03 6.89 4.58 2.31
62 2.31 6.67 4.04 2.63
63 -2.44 1.99 4.77 -2.78
79 2.46 6.56 3.75 2.81
86 -2.05 1.59 3.93 -2.34
88 -2.73 1.71 4.82 -3.11
89 -2.28 2.76 5.36 -2.60
129 3.48 6.34 2.37 3.97
140 2.79 6.60 3.42 3.18
142 3.59 7.00 2.91 4.09
230 2.62 7.00 4.01 2.99
Dependent variable: CHL proficiency
245
These outliers were extreme cases because their ‘CHL proficiency’ could not be well
predicted by ‘Chineseness’ and/or ‘capital’. Among these extreme cases, five candidates (79,
89, 129, 134, and 228) had indicated their interest in participating in an interview. Invitations
were sent to these candidates. Only one respondent (assigned the pseudonym, Adam)
accepted the interview invitation.
At the second stage, Patton’s (1990) maximum variation sampling strategy was applied.
The maximum variation in CHL proficiency was the main concern. The scale score of ‘CHL
proficiency’ of interview candidates that were not extreme cases (30, 46, 49, 59, 66, 77, 87,
91, 92, 120, 139, 141, 225, and 227) was revisited. Their scores presented a wide spread,
from 2.21 (respondent 46) to 6.89 (respondent 77). To capture the maximum variation in their
CHL proficiency, interview invitations were sent to all of these candidates. Four respondents
(Bob, who chose the pseudonym; and Crystal, Dianna, and En-ning, who were assigned the
pseudonyms) agreed to participate in the interview. The scores of their CHL proficiency were
2.21, 2.57, 6.27, and 2.89 respectively. Some of their demographic features also demonstrated
a wide variation in terms of their birthplace, age at which they moved to Australia if born
overseas, and years of formal CHL learning. Taking account of their difference in CHL
proficiency and demographic features, “maximum variation” was considered to be garnered.
Table 6.2 presents an overview of the four respondents, who offered maximum variation (Bob,
Crystal, Dianna, and En-ning), as well as the respondent (Adam), who was identified as an
extreme case.
246
Table 6.2
Overview of participants
Respondent 89 46 87 139 225
Pseudonym Adam Bob Crystal Dianna En-ning
Gender Male Male Female Female Female
Age in years 28 18 18 21 23
Birthplace Indonesia Hong Kong Australia Taiwan Australia
Age of immigration if
born overseas 12 9 months N/A* 13 N/A*
Generation 1st 1
st 2
nd 1
st 2
nd
Resident city Sydney Brisbane Sydney Brisbane Canberra
Citizenship Australian Australian Australian Australian Australian
Language usage at
home
English and
Indonesian with
a few Mandarin
words
English and
Cantonese English
Hokkien and
Mandarin English
Years of formal CHL
learning None 1 5 13 10
CHL proficiency 2.76 2.21 2.57 6.27 2.89
Number of interviews 3 1 1 2 2
N/A*: Born in Australia
6.2 Who were the five participants?
Each participant was interviewed about life experiences associated with their CHL learning
and their Chinese heritage background (see interview schedule in Appendix 4). This section
briefly introduces the five participants and outlines their key characteristics related to this
study. Further thematic analysis of their interview data will be presented in the next section.
Adam: “I just need to catch up!”
Twenty-eight-year-old Adam was born in Indonesia. He moved to live in Australia at the age
of 12. He also lived in Singapore for one year at the age of 21 on an exchange program.
Adam was considered an extreme case because the statistical models could not predict his
CHL proficiency given his investment of various forms of capital. As such, he constituted an
‘outlier’ in the initial quantitative phase. He seemed to be very aware of this, as he agreed, “I
247
think my case is a little bit unique.” This uniqueness may be attributed to his life trajectory.
Adam was born in Jakarta, Indonesia, which he thought was “one of the most challenging
aspects to learning Mandarin”. To understand the challenge that Adam confronted as a CHLL
when growing up in Jakarta, it is useful to provide some background information.
Discrimination against Chinese Indonesians is evident throughout the history of
Indonesia, although more recent government policies have attempted to redress this.
Resentment of Chinese economic success grew in the 1950s as native Indonesian merchants
felt they could not remain competitive. In response, the government approach was to
propagate a stereotype that conglomerates owned by Chinese Indonesians were corrupt. As a
result, the word ‘Chinese’ became synonymous with corruption in the perception of most
Indonesians (Purdey, 2006). The situation deteriorated when General Suharto came into
power in 1967. His approach was to forcibly assimilate Chinese Indonesians, taking
advantage of Chinese economic success whilst eliminating their perceived economic
dominance (Purdey, 2006). In order to do this, he sought a depoliticised system where
formation of a cohesive ethnic Chinese identity was no longer allowed (Purdey, 2006).
Expressions of Chinese culture through language, religion, and traditional festivals were
banned. Chinese Indonesians had to relinquish their Chinese names and ‘happily’ adopt
Indonesian-sounding names. Anti-Chinese sentiment gathered intensity in the 1990s when
major riots broke out in many Indonesian cities. Property and businesses owned by Chinese
Indonesians were targeted by mobs. Many women were sexually assaulted and numerous
people died (Purdey, 2006) . Following these events, large numbers of Chinese Indonesians
fled the country. In the late 1990s, the Asian Financial Crisis had dire consequences for the
Indonesian economy and shook Suharto’s regime. After nation-wide demonstrations against
the presidency organised by university students, Suharto’s 31 years of power came to an end.
248
Subsequent governments launched campaigns to abolish anti-Chinese regulations and rebuild
the confidence of Chinese Indonesians. Although Chinese Indonesians have gradually gained
political and social freedoms as a result of policy reform efforts, anti-Chineseness is still an
issue in Indonesian. It was against this backdrop that Adam was born and brought up before
moving to Australia at the age of 12. He noted, “Because of the challenge, I’ve got a lot of
catching-up to do.”
Bob: “Chinese is part of me.”
Eighteen-year-old Bob was born in Hong Kong and moved to live in Australia when he was
nine months old. He had returned from a two-week family trip in Hong Kong shortly before
his interview. Bob explained, “I just went there with my dad. I haven’t been there for 15 years.
So I went there and we saw our family.”
Bob was the only participant who did not use a single Chinese word during the
interview. Bob’s self-reported CHL proficiency measured in the online questionnaire was the
lowest of the five participants (as shown in Table 6.2). This may be the reason behind his
scant usage of CHL. Even if he did not have many opportunities to use his CHL in his
“younger years”, he intentionally chose Chinese when both Chinese and French lessons were
available in his primary school. In response to the question about the reasons behind this
effort, Bob said, “Because of my Chinese heritage. French would be good but I think it’s just
nothing part of me. But Chinese is part of me. I just think it’s a good opportunity to do that.”
Although he came to live in Australia when he was only nine months old, his embodied
dispositions, or in his own words “part of me”, was durable and transposable. He felt very
comfortable with it. As he said, “You can’t pass Bruce Lee and his movies. He made me feel
happy about my heritage when I was watching these movies and stuff.” He had developed a
249
cultural attachment to the long history of Chinese Kung Fu that contributed to his degree of
comfort with his Chinese heritage.
Crystal: Chinese entertainment stuff is “really funny”.
Eighteen-year-old Crystal was born in Australia. She was a second-generation Chinese
Australian. She said she enjoyed Chinese TV shows. In her own words, “It’s really funny (to
watch Chinese TV shows)!” When asked whether she was trying to improve her Chinese, she
said, “Yes, that’s part of the reason why I watch Chinese TV shows.” For Crystal, watching
Chinese TV shows and learning Chinese seemed to be mutually reinforcing. As she indicated,
watching Chinese TV shows was one way to improve her Chinese proficiency. In return, her
Chinese, to a certain degree, helped her access a wider range of Chinese entertainment. She
reported typing Chinese and searching Chinese websites for Chinese TV shows, which were
not available in English websites, as she said, “You don’t have this kind of thing in Australia.”
When asked whether her CHL benefited her social life in Australia, she said, “I can find my
own entertainment so I can also talk to other people about entertainment. When I go to
Karaoke, I can sing Chinese songs. Yes, I guess it helps me.” Crystal can thus share her
knowledge of Chinese entertainment in a Chinese-speaking community in Australia, where
Chinese popular culture is of common interest. In this situation, CHL proficiency helped her
not only access a wider range of Chinese entertainment, but also build social networks.
Dianna: “I felt how important (it was) to speak my own home language!”
Twenty-one-year-old Dianna was born and brought up in Taiwan. She moved to live in
Australia at the age of 13. When asked to recall her school life in Taiwan, she said she
250
“probably forgot everything”. Interestingly, when asked whether her CHL had ever made her
feel special, she immediately recalled the following experience. Once at lunchtime in high
school, her white Australian classmates were “curious” and asked her and her Chinese friend
whether they could speak Chinese. Dianna said she could because she used Chinese at home.
She then said a few words in Chinese. Her Chinese friend felt “embarrassed” because she was
not able to speak Chinese at all. Dianna recalled, “At that time I felt how important (it was) to
speak my own home language.” She could clearly remember what her white Australian
friends said at that moment, “If you are Asian, you should be able to speak your home
language. If you can’t, it’s sort of a shame.” Dianna considered her CHL a source of “pride”
to her and “part of” her “Chinese identity”. This perception was deeply rooted in herself,
which can be understood as constituting a durable and transposable disposition. As she added,
“I still need to keep my home language going no matter where I am. So I will pass it on to my
children as well.”
En-ning: “Learning Chinese is definitely something I am going to pursue in the rest of
my life.”
Twenty-three-year-old En-ning was born in Australia. She was a second-generation Chinese
Australian. At the time of the interview, she was attending Chinese programs at a prestigious
university in China. She reported that she did not see any point in learning Chinese when she
was a child, growing up in a “very white area”. This implied that she did not have many
opportunities to use Chinese at that time and she was not interested in learning the language
at all. However, she said she did not know what “came over” her in Year Nine when she
decided to learn Chinese in school. From then on, she started to realise that learning Chinese
was “an incredible personal experience” for her and she was trying to do “as much as” she
251
could in Chinese. In return, En-ning profited from her CHL learning. Her CHL proficiency
helped her win a prestigious scholarship at a leading Australian university and a national
scholarship awarded by the Australian Federal Government. She confessed, “Of course I
don’t think I would have got the scholarships if I didn’t have at least some Chinese language
background.” En-ning spent a gap year in China, which “totally changed” her life. She was
able to “explore the other side”, or “the Chinese side” of herself. She said, “I feel a lot
happier, especially in my gap year when I had the opportunity to explore that side of who I
am.” Learning Chinese was so “amazing and incredible” for En-ning that she said, “Learning
Chinese is definitely something that I am going to pursue in the rest of my life.”
6.3 What did CHL mean to the five participants?
Different reasons and various meanings behind CHL learning emerged from the above
glimpse of the five participants. When legitimised in given fields, participants’ CHL
competence could serve as linguistic capital. The exchange value of this linguistic capital can
produce benefits or profits in these fields. These benefits and profits can be understood as a
return on investment in CHL learning. The quantity and quality of production was dependent
upon the structure of the participants’ lived social worlds at a given time. As such, CHL was
reported to have rich and varied meanings for all participants. These meanings emerging from
the interview data resonated with Bourdieu’s theorisation of habitus, capital, and language
choice within a given field. The data are reported, interpreted, and analysed thematically in
the following sections.
252
6.3.1 Potential production of economic capital through CHL learning
When asked whether their Chinese language could contribute to job opportunities in future,
all participants indicated that one of the benefits from learning Chinese was the expectation
of more and better job opportunities. Dianna shared a story about one of her Chinese friends
who got a job as a translator in a company because of his Chinese proficiency. Dianna
seemed to cast no doubt on the value of Chinese competence in the labour market. She
described always telling her friend’s story to her brother and saying, “You will be thankful
when you get a job because of your ability to speak two languages.” She explained, “When
you are looking for a job, people tend to consider someone who can speak two languages or
more than two. You will find yourself getting that priority.” By saying this, she indicated her
belief that her Chinese competence would be valued in such a field where people considered
both Chinese and English important. When valued in a future job market, her CHL
competence will serve as linguistic capital by which she can “get that priority”.
This view was shared by all participants. Adam, who was doing research in air
pollution, hoped that his Chinese competence would open up opportunities to find work to do
with air pollution in China. Bob, who was studying paramedical science, reported that his
Chinese competence might be helpful if he wanted to get a job in Hong Kong as a paramedic.
Crystal, who wanted to possibly work in Hong Kong one day, considered that her Chinese
competence might function as an instrumental advantage for her. En-ning, who changed her
major from arts and law to Asian studies and law, considered her Chinese competence useful
if she wanted to “pursue Australia-China relations or work as a lawyer” in her future career.
In summary, all participants reported if valued in a future labour market, or a given
field in Bourdieu’s sense, their CHL competence would become a valuable asset for them.
These future labour markets where Chinese is valued will become extra fields open to these
253
participants. This accords with the findings indicated by previous studies (Francis et al., 2009;
Weger-Guntharp, 2006; Wong & Xiao, 2010) that CHL competence is explicitly associated
with extra job opportunities, and ultimately convertible into economic capital.
6.3.2 Production of cultural capital through CHL learning
When asked to explain their Chinese learning experiences and practices, Adam said that he
bought some Chinese textbooks and worked through them by himself; Crystal reported that
she bought some Chinese TV series and watched them as a way of learning Chinese; and
Dianna indicated that she collected some Chinese books for reading and family discussion.
These Chinese textbooks, Chinese TV series, and Chinese literature are material cultural
goods that can be understood as objectified cultural capital produced through CHL learning.
When asked to talk about language and culture, Adam explained, “I have a bit of a
philosophical aspect in learning a language or any languages because it provides a wider
window to look into the culture and see how people react to things.” This philosophy made
sense when he reported that he started to understand local people’s particular ways of doing
things during his Chinese language study tour in Shanghai. These understandings gained
through CHL learning became an embodied asset. As Bourdieu (1991) explained, embodied
cultural capital would include the know-how capacity that people bring with them when
moving across different social spaces. By virtue of these embodied resources, when Adam
accompanied his sister to visit Shanghai, he became the person “who showed her around”,
told her “how unique people in Shanghai are”, and “trained her how to cross the road safely”.
Apart from his capture of the local ways of doing things, Adam also reported gaining
some Chinese cultural knowledge through his Chinese learning. He explained:
254
We can use the word ‘networking’, but it’s not as great as the Mandarin word, (I am
sorry for my tones), ‘关系’. In China, ‘关系’ is a lot deeper concept. It’s more about
knowing the right people and somehow having a working relation. It’s like a ‘you
scratch my back and I scratch your back’ thing.
This excerpt demonstrated his understanding of the special meaning of the Chinese word ‘关
系’ , equivalent to the English word ‘networking’, within a Chinese cultural context. He
gained this Chinese cultural knowledge through his CHL learning. Later he added:
I am always fascinated with the old Chinese sayings, which still ring true to this day. It
just has a deeper meaning when you say it in Mandarin or even in classical Chinese
instead of reading the English translated one.
He gave a few examples of the old Chinese sayings and said, “Remember the wise sayings
from our forefathers: 一山还有一山高 (there is not a mountain which is the tallest); 天外
有天 (there is always another heaven beyond this one).” These sayings indicate that no one
can be the best and people should be humble. It was apparent that Adam appreciated the
meaning behind these sayings. He referenced these sayings when trying to argue that he
would not feel too proud of his Chinese competence, though he was happy with his progress
in Chinese. Through CHL learning, Adam reported that he had gained Chinese cultural
knowledge and embedded this knowledge into his mind and body, from which his
behavioural patterns, actions, and perceptions emerged.
Bob also reported gaining some Chinese cultural knowledge through his Chinese
learning. When asked to talk about language and culture, he reported that his mum taught him
255
something about “the formation of Chinese characters” and “how different characters had
changed over the years.” It was his mum’s inculcation through CHL practice that generated
this aspect of his Chinese cultural knowledge. Similarly, Crystal reported that she really
enjoyed Chinese entertainment, especially Chinese TV series, and she attributed this taste to
her upbringing, as she explained:
I grew up watching that. When I was very little, my dad introduced me to these series.
My parents never formally taught me Chinese really. But they show me TV shows. We
used to watch together. That’s what I did.
As Bourdieu (1973) argues, the reproduction of cultural capital emphasises the importance of
the early familial environment for the learning and accumulation of cultural capital. By this
he meant that reproduction of cultural capital is not established in a vacuum but is
intergenerationally sponsored. This intergenerational influence urges each agent to reproduce
the cultural capital of his or her group, augmenting it if possible. Through CHL learning,
Bob’s mum and Crystal’s dad were able to pass on the Chinese cultural knowledge to them as
the reproduction of embodied cultural capital. When asked whether her Chinese language
benefited her, Crystal said:
Yes, I guess so. It exposes me to more entertainment stuff. Before, when I was little, I
used to watch series but my dad had to give me the series to watch. Now I can go
online and type in Chinese to find my own series to watch.
256
By saying “more”, Crystal indicated that she would not have been able to access Chinese
entertainment resources without her CHL learning experience and CHL competence. When
she was little, her dad had to mediate access to these resources for her because her Chinese
was too limited. When she grew up, she was able to use her Chinese to search Chinese
websites for “more entertainment stuff”. Moreover, both Crystal and En-ning reported that
learning Chinese enabled them to sing Chinese songs. For Adam, Bob, Crystal, and En-ning,
their Chinese cultural capability generated through learning CHL can be understood as
embodied cultural capital that helped them access a wider range of Chinese cultural
resources.
In addition to objectified and embodied cultural capital acquired through CHL learning,
CHL competence was also reported to produce institutionalised cultural capital. All
participants except Adam reported that they had undertaken formal Chinese learning in
schools or at universities.
When I was probably in Grade Three or Grade Four, my parents sent me to a Chinese
school to learn Mandarin… In Grade Seven…Chinese was the thing in the curriculum
(of my school). We got to learn Chinese. (Bob)
I think I took my first formal Chinese lesson when I was in Year Three at my primary
school… I did two years of Chinese at high school in Year Seven and Year Eight…
From Year Nine to Year 11, I went to my community Chinese school, run by Chinese
people… I am going to do my first Chinese course this coming semester (at my
university). (Crystal)
257
I went to one (Chinese community school) in Sunnybank for a few months. (Dianna)
I went to Cleveland Street Saturday Chinese School until…possibly Year Six. And
then...my school started offering Chinese. And then I did it in Year Nine… I did
Chinese Five, Chinese Six, and Chinese Seven in my first semester and the second
semester and the second year first semester (at my university)… And then I went for an
exchange in Santa Barbara in California at this time last year for a semester. I took like
three Chinese language classes as part of my exchange requirements. (En-ning)
Because the Chinese courses were part of participants’ (except Adam) curriculum,
participants (except Adam) received credits towards the programs that they were enrolled in.
These credits can be understood as institutionalised cultural capital.
6.3.3 Production of social capital through CHL learning
When asked about their language choices at home, at work, with friends, and in their social
lives, all participants indicated that Chinese was useful and beneficial in at least one of these
settings. First of all, they all indicated that their Chinese competence helped strengthen their
family ties. By their accounts, this was particularly important when some family members did
not have, or did not have enough, English competence. Dianna reported that Chinese usage at
home was the only way to communicate with her parents because they did not have any
English. Crystal and En-ning reported that Chinese usage helped them communicate “more”,
particularly with their grandparents:
258
I would like to communicate with my family a bit more. I really like my more distant
family (grandparents), so I would like to be able to talk to them more when I visit them
and tell them about my life. (Crystal)
It was very hard to communicate with them (grandparents) before I properly learnt
Chinese. I guess it’s another part of why learning Chinese is a personal thing because I
want to better speak to my grandparents. (En-ning)
In the excerpts above, Crystal and En-ning used the words “more” and “better”. By this they
meant something they did not have or did not have enough of before. The connotation
embedded in the words “more” and “better” resonates with the capital metaphor because
capital can accumulate and generate ‘added value’. For these participants, the added value
generated through CHL learning was better intergenerational communication, by which their
family ties could be strengthened.
Secondly, CHL competence could reportedly build friendship networks. All
participants talked of speaking Chinese with their friends, either in Australia or in China, or
even in other parts of the world. When asked about their language choices with friends,
Crystal said, “I also have some mainland China friends whom I like to talk to in Chinese
sometimes.” In this respect, CHL is important for Crystal to maintain and strengthen the ties
with her friends in the Chinese Mainland. When asked the same question, Dianna explained:
I do have friends from Hong Kong, or China. We do speak Chinese to each other, but
not in class to be polite. If you meet someone with an Asian background who can speak
Chinese, you always feel comfortable to speak Chinese to them.
259
Later she added, “I tend to be with Asians (Chinese) more, so we can speak our home
language, Mandarin Chinese. Sometimes we can speak about things that we can’t describe in
English. We are comfortable that way.” Her friendship could be strengthened through this
more comfortable way of communication in Chinese, especially when there was something
that they could not “describe” in English. In such a situation, CHL had a unique and
instrumental function to facilitate communication between Dianna and her Chinese friends,
which they could not achieve through any other language. En-ning also reported how her
Chinese competence benefited her friendships:
I think of how many diverse and amazing experiences I have been able to have as a
result of coming to China and learning Chinese. For example, I have memories of
overseas Chinese from all over the world hanging out in a dorm room and sharing
stories about our experiences as overseas Chinese…Chatting to a bunch of Chinese
girls on a train, you know anything that is relevant to their lives. I think if I can’t speak
some Chinese, I would not have access to it.
Moreover, CHL competence by the participants’ accounts benefited their daily lives.
All participants described experiences when their Chinese competence helped them socialise
with other people. Adam used to mix with other Chinese-speaking people in a Chinese mall
when he was in Singapore. When asked in the second interview why he wanted to ‘catch up’
with his Chinese, Adam added:
I want to catch up with my Chinese language so that I can converse with anybody easier.
I think I am sort of on my way there. For example, I just helped a couple of my friends
260
with their graduation and most of them are Malaysian. Even a year ago, in the previous
graduation season, I wouldn’t even think of being able to talk with any of their parents
in Chinese. This year at least I could say a few things.
As a return on his efforts in CHL learning, Adam was able to socialise in a more active way.
Bob and Crystal reported the experience at work of using Chinese to help people who could
not speak English. Crystal was also very happy that she was able to use Chinese in China
when shopping. Dianna taught small children simple Chinese keywords or sang Chinese
songs with them over the Chinese New Year. En-ning was studying in China when the
interviews with her were conducted. She indicated that her Chinese language competence
was indispensable because she had to use Chinese in her daily life to socialise with other
Chinese people.
In contrast, a lack of CHL proficiency was reported to hinder the establishment and
development of social networks. When asked whether it was a struggle to speak Chinese, Bob
said:
Definitely I was keen to socialise with people but it’s hard because I don’t know what
to say. If I knew what to say, I would definitely say (something). Yes in a way, it was a
struggle. But I was willing to try and make that effort to say something.
On the one hand, a lack of CHL proficiency precluded Bob from socialising with
Chinese-speaking people. Without certain level of CHL proficiency, it was difficult for Bob
to establish social networks with Chinese-speaking people. On the other hand, Bob described
trying his best to practice his CHL in order to better enable socialising.
261
In brief, CHL competence has maintained, extended, or improved the social networks
for all participants in their Chinese-speaking social world, in terms of tightening family ties,
developing new friendships or improving their existing ones, and socialising with other
people. This is consistent with the existing literature. As indicated in previous studies
(Francis et al., 2009; Wong & Xiao, 2010; J. Zhang, 2009), CHL as a means of
communication has a functional significance in terms of establishing valuable networks
within fields where Chinese is the medium of social relations.
6.3.4 Production of symbolic capital through CHL learning
CHL competence reportedly became symbolic capital when recognised or valued by others.
Adam recalled his feeling that local people understood him when he was in Shanghai:
I felt really great! It was just like a little kid managing to do something different for the
first time. I had thought I would have never done it because of all the difficulties that I
had. For them to be able to reply to me: “I understand what you said”, it looked like I
managed to climb up a pretty high mountain and I would be able to do it. The feeling
was like excitement and wanting to do more along the way…Yes, it really made me
think, “I can do this. Let’s learn more.”
This “excitement” and motivation to “learn more” may be attributed to his experience that his
Chinese language was understood, accepted, and legitimised by the local Shanghai people.
Other participants reported similar experiences of recognition and affirmation when their
Chinese language was recognised or valued by others. When asked what other people thought
of his Chinese ability, Bob recalled, “My aunty was pretty nice. She said that I am good for
262
someone who has been brought up in Australia.” When asked about her reasons for learning
Chinese, Crystal said that her friends’ parents made a fuss of her when she talked to them in
Chinese. She recalled, “They kind of like me more.” Dianna had the experience that she was
praised by her parents’ friends for her CHL competence. She recalled what her parents’
friends said to her, “Oh, it’s really good you can still speak Chinese here”. She also recalled
the experience that she was asked by her white Australian classmates in high school to speak
Chinese. When her classmate said, “I only speak English. You can speak two languages”,
Dianna reported that, “I was really, really proud of myself. I have two languages. I don’t feel
any shame at all.” Because her Chinese proficiency was recognised by others, she did not
“feel any shame at all” for being able to speak Chinese. Instead, it was a source of pride for
her. En-ning recalled, “In terms of my grandparents, they are very proud and happy that I am
pursuing this language.” When asked whether learning Chinese has ever helped her win any
qualifications or awards, En-ning was “pretty sure” that “almost everything” she was able to
win was “partially because of” her Chinese learning efforts. These included a prestigious
scholarship at a leading Australian university and a national scholarship awarded by the
Australian Federal Government. As she explained, “Of course I don’t think I would have got
the scholarships if I didn’t have at least some Chinese language background.” Because of the
symbolic value attached to her Chinese language, En-ning said, “I am very aware how lucky I
am to be an Australian with a Chinese heritage.”
For these participants, CHL competence became linguistic capital with legitimised and
symbolic value, which was then convertible into symbolic capital. As indicated in previous
studies (Francis et al., 2009; J. Zhang, 2009), CHL competence helped CHLLs gain praise,
respect, pride, status, and honour.
263
6.3.5 Making sense of the habitus of ‘Chineseness’ through CHL learning
As argued in Chapter Three, habitus is rooted in cultural history, or “previous state” in
Bourdieu’s term (2000, p. 161). For some participants, the origins of their ‘Chineseness’
reportedly carried deep meaning and relevance through their Chinese learning. En-ning took
the opportunity when she was learning Chinese in China to explore her family origin. She
said, “This 春节 (Chinese New Year) I found my grandparents’ 老家 (hometown) in the
south. I spent two weeks there.” It should be noted that the Chinese word ‘老家’ has a
different meaning from the English word ‘hometown’. ‘老家’ is the ‘root’ of a clan, a native
place where one’s ancestors originated and were buried. It signifies not only the physical
place itself but also the genealogical connections and the deep attachments to the land,
customs, and compatriots, forged through generations of shared ancestry, history, culture, and
language. En-ning seemed to attribute this trip to explore her cultural roots to her growing
Chinese competence. As she explained, “I wouldn’t have managed if I didn’t have at least
some Mandarin. Even my parents have never been (there). It was incredible.” En-ning’s
account is in line with the literature that the heritage background nurtures the HL learning
efforts and reinforces HLLs’ identity (Carreira, 2004).
Adam thought that Chinese language resonated with ‘Chineseness’, as he explained,
“For some people, they speak Chinese so that they can claim their ‘Chineseness’”. When
asked about the driving force behind all his efforts in learning Chinese, Adam replied:
The driving force is that I couldn’t do it because of the political reasons. I could keep
saying I was the victim of the political situation back then, but now I have the chance
and I will not let the opportunity pass. I am sort of trying to discover my heritage again.
It was denied but now I have the chance and I am going to grab it.
264
He concluded his first interview with the following comments:
I am not a typical Chinese because most of the time, let’s say Westerners, maybe use
that word, they probably learn Chinese because they have the opportunity and they
have always been able to do it. Well, in my case, I was just a little bit behind because of
the policy I had back then. But of course I can’t say, “Hey, it got banned. I got left
behind. There is no point of learning.” No, that’s not the case. I just need to catch up.
That’s all.
Adam was trying to explain his uniqueness here. In his perception, “Westerners” can “have
the opportunity” to learn Chinese and “they have always been able to do it”. In contrast, he
did not have the opportunity and he was not able to do it while growing up in Jakarta. Here he
might use the terms “Westerners” and “typical Chinese” to refer to white Australians and
Chinese Australians respectively, who were brought up with the opportunity of learning
Chinese. In fact, he did indicate earlier in the interview that the multicultural society offered
Australians the opportunity to learn Asian languages because these languages were “just
around them”. Compared with these “Westerners” and “typical Chinese”, he felt left behind
in terms of Chinese learning. This can explain his lower Chinese proficiency reported in the
quantitative phase. This was also the reason that he wanted to “catch up”. When asked what
exactly he wanted to catch up, Adam replied, “I want to catch up with my Chinese language.”
By this “catch-up”, he explained that he wanted to “discover” his “heritage again”. By saying
“again”, he indicated that he sought to recapture his claim to ‘Chineseness’ that “was denied”
in the past due to the difficult political situation. He added this comment at the end of the
third interview:
265
For me, Mandarin is a tool to enforce the Chinese identity…It’s to satisfy my curiosity
about my heritage and how history, culture, etc. have shaped who I am today as a part
of my own ‘who do you think you are’ project.
It is worth nothing that Adam was referring to the TV genealogy program ‘Who Do You
Think You Are?’ It is an Australian television documentary series, following the BBC series
of the same name. Each episode profiles a celebrity tracing his/her family roots. Adam
referred to this program due to its promotion of genealogy, which resonated with his
“curiosity” about his ‘Chineseness’.
Other participants also claimed that Chinese language carried deeply rooted meanings
for them. When asked why he chose to learn Chinese in school, Bob said, “Chinese is part of
me.” When asked what Chinese language meant to her, Dianna replied, “It’s also part of my
identity, my Chinese identity. I was born into this culture, this colour, and this language, so
it’s part of my identity, my Chinese identity.” When asked, “Have you ever thought about
why they (your parents) pushed you to learn Chinese”, En-ning replied:
Yes, I have thought about it. I guess they want me to have a connection with that side of
who I am…They also gave us (me and my brother) opportunity to explore the other
side, the Chinese side of us when we were very young.
En-ning’s parents hoped to reproduce a sense of ‘Chineseness’ in their children through
learning CHL. En-ning seemed very engaged in this reproduction of ‘Chineseness’. She spent
her gap year in China learning Chinese, which “totally changed” her life. She said, “I feel a
lot happier, especially in my gap year when I had the opportunity to explore that side of who I
266
am.” When asked at the end of the interview to add some more comments to her Chinese
language learning and Chinese background, she said:
I am so glad that I discovered this so early in my life. I do meet a lot of overseas
Chinese who came to China much older, looking for something. But I am glad I came
early…I think if we have the opportunity to explore this other side of who we are or
might be, that can be very empowering, which in turn can very much affect our life
experiences.
In short, participants’ CHL learning and practices resonated with a constitutive dimension and
sense of themselves. As such, CHL learning and practices helped recapture or reinforce a
habitus of ‘Chineseness’ for the participants, which was embraced as “part of” themselves or
“the other side of” themselves. This is in line with the existing studies where CHLLs reported
that Chinese was a central part of themselves (Comanaru & Noels, 2009).
All participants indicated that they intended to reproduce ‘Chineseness’ in the future
through intergenerational Chinese learning and practice. When asked, “If you were to have
children, would you like to encourage them to learn another language other than English?” all
participants agreed that they would like their children to learn Chinese.
I would love to see them learn Mandarin. Yes, I would love to see them learn
Mandarin…It’s definitely biased of course. (Adam)
I wouldn’t force them. They can decide what they want to know, like where their
grandparents’ heritage was from. But I won’t force them. If they choose (Chinese) and
267
they are interested, by all means I will support them and try to coach them…They can
learn where their dad is from. (Bob)
I would like my kids to learn Chinese…so that we don’t lose it as time goes on. We are
Chinese family and we should speak Chinese. (Crystal)
I need to keep my home language going no matter where I am. So I will pass on to my
children as well. (Dianna)
I would encourage them to learn Chinese…I personally would like them to learn
Chinese because of the personal connections they would have. (En-ning)
Dianna expressed the strongest sense of the maintenance of her CHL and definitely intended
to pass it on to her children. This disposition may be attributed to her upbringing in Taiwan,
where in general Chinese language and culture are more valued than in the places where the
other interview participants were brought up. This conforms to Bourdieu’s contention that
‘field’ generates ‘habitus’. Adam’s “biased” disposition demonstrated his eagerness to claim
his ‘Chineseness’ and transfer it to his children through CHL learning because their current
lived world, Australia, does not ban their CHL and ‘Chineseness’. Bob, Crystal, and En-ning
also demonstrated a certain degree of desire to maintain the ‘Chineseness’ in their next
generation through CHL learning because they have all enjoyed the benefits of CHL to some
extent.
In brief, the habitus of ‘Chineseness’ was explicitly linked to past roots, present
moments, and future anticipations. This habitus of ‘Chineseness’, as a system of internalised
268
cognitive and motivating structures (Bourdieu, 1977b, 1990), is produced by structures of
past and present cultural and social environment and will be reproduced in future CHL
learning through its generativity. When CHL is legitimised and imbued with symbolic value,
the loyalty to CHL may persist over many generations of Chinese Australians who relearn
CHL that symbolises their ethnic identity, and in Bourdieu’s sense, their habitus. This CHL
learning is a strategy to reproduce the habitus of ‘Chineseness’ in the next generation and
even generations further removed. A habitus of ‘Chineseness’, to borrow a Bourdieusian
metaphor, contains the genetic information which both allows and disposes successive
generations to reproduce the world that they inherit from their previous generation (Bourdieu
& Passeron, 1990). In this respect, this durable and transposable ‘Chineseness’ is transmitted
across generations and over historical time through CHL learning.
6.3.6 Capital in different fields
As argued in Chapter Three, capital only accrues value in a particular field. The interview
data illustrated this well. Adam has lived in three different countries, where Chinese language
did not carry the same value. He was born in Jakarta, Indonesia. When he grew up there,
learning Chinese was almost impossible. He recalled and revealed this experience as follows:
During the Suharto era they had a bit of a policy where nobody could learn any other
language other than English or Indonesian. You can say that the Mandarin language was
banned…Anything with Chinese writing wasn’t allowed. Interestingly, if you go to
Indonesia around that period of time, filling the customs card, like what we fill when
entering into Australia, there was actually a part where it says whether you have
anything that has Chinese writing or something like that.
269
During the Suharto era in Indonesia, English and Indonesian were positioned at the top of the
language hierarchy and therefore were considered legitimate and valuable linguistic capital
within this particular field. Commitment to learning these two languages was an astute
investment, given the return on which language learners could have access to resources, such
as better education, more job opportunities, and a better life. In contrast, Chinese language
was illegal and would have even accrued a penalty. Investment in Chinese would not have
any positive public returns. This was “one of the most challenging aspects to learn Mandarin”
for Adam. The situation was very different when Adam lived in another field, Singapore,
where Chinese was widely used. When asked about his Chinese practices in Singapore,
Adam recalled:
What happens is that they have more textbooks that teach little kids Mandarin. I just
bought them and I actually went through them myself from the first grade to the second
grade…I learned quite a lot of Mandarin in Singapore because more people can speak
Mandarin. I was just looking at the little kids around to see what words they used. If the
little kids use the words, I can bet ten bucks that it’s going to be one of those few words
that I need to know.
In the second interview, Adam added, “It looks like they (Singaporeans) put a lot of
importance on our ability to speak Chinese.” In his perception, Chinese language seemed to
carry much value in Singapore. This can explain why he said that the reason to do the
exchange in Singapore was to “try to learn a bit more Mandarin along the way”. This one
year in Singapore helped his Chinese, as he added, “When I was in Singapore, I definitely
learned more Mandarin. I stayed there for one year and I took my time learning it… At least
270
the writing and the listening definitely improved when I was in Singapore.” However, Adam
had not pursued any formal Chinese learning in Australia. This may be because the social
fields he lived in did not accord similar value to Chinese. When asked whether many students
chose to learn Chinese and whether Chinese was popular among them, he recalled:
I would say nearly none. Nobody talked about Chinese at all. Maybe because the school
did not offer it, nobody thought about it at all. Even if I asked other people if they were
interested in Chinese, they would say “no, what for?”
When asked about his language choices in Australia, Adam explained his language choices at
home, at work, with his friends, and in his social life. At home, he spoke mainly Indonesian
with his parents who “put more effort into (his) English because it is pretty much the working
language of the world”. He spoke half English and half Indonesian with his sister. At his work,
he believed that his Chinese competence may bring him opportunities in his future career, but
he added, “I wouldn’t say it will be a direct benefit. For example, because I am a researcher,
probably…especially in science, everybody just uses English.” He spoke mostly English with
his friends, who tended to say, “Yeah, Mandarin is important but not that essential.” In his
social life, he did not have many chances to use Chinese because “everybody here can speak
English anyway”. For Adam, Chinese was not a valuable resource in Australia.
According to Adam’s experiences, Chinese accrued different values in Indonesia,
Singapore, and Australia. Likewise, Chinese was valued differently in Hong Kong and
Australia. Because of the colonial history, English was once the only official language in
Hong Kong. Since the ‘handover’, English has kept its value due to this history and its global
utility while Chinese has gained more legitimised value due to the current political situation.
271
Both English and Chinese serve as official languages and linguistic capital in Hong Kong.
They have both been positioned at the top of the language hierarchy. Bob recalled his
experience during his recent Hong Kong trip:
My dad’s sister said, “It’s very important you should learn and you should try to
improve your speaking, so you can communicate well”…They are quite supportive.
When they knew that I was trying to say something, they would try to help me out to
get the point cross.
The words of Bob’s aunty implied that Chinese accrues value in Hong Kong because “it’s
very important”. Unlike Hong Kong, Bob perceives that Australia is not a place where
Chinese is valued:
Well, when I was there (Hong Kong), I thought it (Chinese) was a quite important thing.
But since, you know, I have been here (Australia) for so long, I felt in my younger
years it wasn’t quite a benefit for me because I don’t use that language.
This experience was shared by En-ning. She reported that she “grew up in a very white area”
and indicated she did not have many chances to use Chinese when she was small. When
asked about his language choices in Australia, Bob explained that English was his dominant
language at home, at work, and in his social life. However, for Bob, Chinese did have its
value in other fields. For example, when asked why he chose Chinese in primary school,
apart from the reason that Chinese was “part of” him, Bob replied, “I think a few of my
friends also did that”, suggesting that learning Chinese was a popular choice among his
272
friends. In order to better maintain his membership in this friend network, or this particular
field in Bourdieu’s sense, where Chinese had a degree of legitimate value, Bob pushed
himself to invest in his CHL learning.
Similarly, although English was the dominant language for Crystal, her Chinese
competence was sometimes valued in certain fields. She said, “I have some Chinese friends.
Sometimes I think it is good when I go to their house and I talk to their parents in Chinese,
they kind of like me more.” For Dianna, her Chinese accrued value when she translated for
her parents in given situations. When asked why she considered her Chinese competence a
source of “pride”, she responded, “I guess every time I bring my parents to places like
immigration, post office, and Brisbane City Council, where they need translation, I am the
translator. I feel proud of being able to translate for my parents.” In these situations, Chinese
language became important capital for Dianna.
In contrast to the linguistic fields in Indonesia, Singapore, Hong Kong, and Australia,
the linguistic field in the Chinese Mainland positions Chinese at the top of the language
hierarchy of all the 129 languages in China (Sun, Hu, & Huang, 2007). Within the Chinese
language, Mandarin, the largest group of Chinese dialects in terms of population and
geographical distribution, is considered the most valuable linguistic capital, the possession of
which can determine people’s position in a given field. Different geographical regions have
their own legitimised Mandarin accents. Adam reported an example for this. When he was in
Shanghai, speaking to local people with the wrong tones, the locals said, “Hang on, how
come you can’t speak Mandarin! Where are you from?” Likewise, En-ning recalled a similar
story during her study in Beijing. “I am definitely conscious sometimes like I’ve been to a 食
堂 (dining hall) with my friends and people will be looking at me because they are all talking
Chinese but they can tell I have a very funny accent.” In such a social field, the quantity and
273
quality of the pertinent linguistic capital, that is, a particular accent of Mandarin, can
distinguish the ‘in-group’ from the ‘out-group’. It was the absence of this linguistic capital
that caused Adam and En-ning to feel excluded from a legitimate Chinese cultural citizenship.
This cultural citizenship can be understood as a set of “cultural practices and beliefs” and
“hegemonic forms that establish the criteria of belonging within a national population and
territory” (Ong et al., 1996, p. 738). In this case, the different accent of Adam and En-ning
did not fit the “hegemonic forms” within the Chinese cultural and linguistic fields and
therefore failed to satisfy the criteria of being a legitimate Chinese cultural citizen.
Drawing insights from these data, different political states and social places, such as
family, friendship network, workplace, and school, function as fields in which CHL
competence is deemed necessary or unnecessary, and simultaneously, as fields in which the
value of CHL competence is determined. The more legitimate CHL competence is in a given
field, the more necessary and profitable it is to be competent in it, and the more damaging
and costly it is to be incompetent. For Bourdieu (1989a), the social “sites” are “markets
which, by their positive or negative sanctions, evaluate performance, reinforcing what is
acceptable, discouraging what is not, condemning valueless dispositions to extinction” (p. 85).
In other words, CHL competence never accrues absolute, universal, or guaranteed value.
Instead, its value conforms to and depends on certain logic, interest, and rules of exchange in
a market system. The competence of CHL as linguistic capital realises its exchange value
through its usage in a particular way and in a given field.
6.3.7 Catering to the field of forces
In Confucian social worlds, standards of excellence and successful learning may be
determined both by the individual learners and by “significant others, the family, the group,
274
or the society as a whole” (K. S. Yang, 1986, p. 114). This view was shared by the current
study. What emerged from the data was that participants’ CHL practices were shaped by the
forces within particular fields. These forces established rules of the fields that structured
participants’ agency and shaped their practices in line with these rules. For example, within
the field of family, participants reported the ‘rule’ of parents’ forcing their children to learn
Chinese. Bob recalled:
When I was probably in Grade Three or Grade Four, my parents sent me to a Chinese
school to learn Mandarin. I wasn’t very interested in that. I said, “I am not interested in
this. There is no point sending me to do this.”
En-ning recalled similar experiences:
Basically until I was in Year Nine, I hated learning Chinese with a passion. I think it’s
probably the thing I hated the most in my life. It was definitely my parents forcing me
and my brother to go. It was not my own choice. From my memory, I used to ask to
quit all the time.
Likewise, Dianna shared a friend’s story:
One of my friends actually told me that…he was born here (in Australia) and his mum
forced him to learn Chinese when he was small. He hated it. He didn’t want to go and
he thought it was the hardest language ever.
275
By these accounts, Bob, En-ning, and Dianna’s friend were not willing to learn Chinese when
they were small. Instead, they were commanded and pressured by their parents to do so. This
resonates with previous studies (Lao, 2004; Weger-Guntharp, 2006; J. Zhang, 2009): CHLLs,
when small, had to obey the parental power, which functioned as the rule in the field of
family.
Apart from parents’ forcing children to learn Chinese, En-ning reported a rule within
the field of the Chinese language classroom. She recalled:
I think there are often different expectations and standards to which Chinese
Australians are held. I’ve felt this throughout my experience of learning Chinese and
just this week, I had one of my Chinese teachers tell me that she holds me to a higher
standard, expects more of me, etc…I personally think that these different standards are
what drive me to keep learning. That my teachers often expect more of me gives me a
lot of motivation.
Teacher’s expectations and standards were reportedly imposed on En-ning, because of the
Confucian rule that teacher-student relations are strongly hierarchical (Biggs & Watkins,
1996). Playing by these rules within the field of classroom, willingly or unwillingly, En-ning
tended to put more effort into learning Chinese.
In addition, the stereotypical perception that looking Chinese meant being able to speak
Chinese was reportedly another driving force behind participants’ commitment to learning
Chinese. By their interview accounts, Adam, En-ning, and Dianna’s friend were all
confronted with the question: “Why can’t you speak Chinese while you look Chinese?” This
rule produced embarrassment, discomfort, and contradictions within themselves. As such,
276
they were in a disadvantaged place within social fields where this stereotype functioned as a
positioning rule. These social fields then became a site of struggle for them and CHL
competence became more valuable capital for them in order to play within the rules of these
fields.
6.4 Are the findings consistent?
The previous section reported the main themes that emerged from the interview data. CHL
reportedly had rich and varied meanings for all participants: CHL had the potential to
produce economic capital; participants gained a wider range of cultural, social, and symbolic
capital, and reinforced or recaptured their habitus of ‘Chineseness’ through CHL learning; at
the same time, these different forms of capital accrued differential value in different fields;
and participants’ CHL learning was shaped by the structures of these fields. In addition, the
interview data supported the findings from the initial quantitative phase, which concluded
that participants’ capture of habitus of ‘Chineseness’ and their investment of various forms of
capital in CHL practices contribute to their CHL proficiency. In Bourdieu’s sense, the
production of language competence demanded capital and habitus. In the subsequent
qualitative phase, learning CHL was reportedly both conscious and unconscious practice for
the interview participants. The production of CHL competence through participants’
intentional and unintentional learning will be discussed in this section.
6.4.1 Capital and CHL proficiency
It will be recalled that the initial quantitative phase justified the positive contribution of
various forms of capital to CHL proficiency. In the subsequent qualitative phase, the
277
interview data indicated that CHL competence was reportedly the return on participants’
investment of economic, cultural, social, and symbolic capital.
Firstly, the investment of economic capital helped improve participants’ CHL
proficiency. Adam bought some Chinese textbooks in Singapore and went through them by
himself. The money spent on textbooks seemed worthwhile. This helped him learn some
simplified Chinese characters because the characters he had learnt were the traditional ones.
He recalled his first response to reading these textbooks: “At first it’s a bit odd. I thought
hang on, what is it? Oh, hang on, it’s the simplified character.” Besides the money spent on
Chinese textbooks, he paid for the air tickets for the Shanghai study tour. He said, “The only
thing (I had to pay) was the air ticket” because the local expenses of the study tour were
covered by the program. The money spent on this trip seemed worthwhile too. As he recalled,
“The one that really, really made me learn Mandarin was that 2010 cultural tour…It was such
a great experience…2010, as you can see, was a big eye open-up.” Like Adam, Crystal also
spent some money learning Chinese. She bought Chinese TV shows and watched them to
improve her Chinese. When the interview was conducted, En-ning was doing a fee-paying
program at a leading Chinese University in Beijing. When asked, “Do you think it’s worth it”,
she replied, “Yes, it’s worth it.” Given the intensive training, En-ning expected to improve
her Chinese proficiency. As she confessed, “I am very comfortable saying that 95% of the
Chinese I have learnt has been in China.” For Adam, Crystal, and En-ning, the money spent
on Chinese learning can be understood as an investment of economic capital eventually
convertible into their CHL proficiency.
Secondly, the investment of cultural capital reportedly contributed to participants’ CHL
learning. When asked, “Are you still trying to improve your Chinese language”, Crystal
replied, “Yes, that’s part of the reason why I watch Chinese TV shows.” As stated earlier,
278
Crystal either bought or downloaded Chinese TV shows. These Chinese TV shows, or
cultural goods, demanded time and Chinese cultural knowledge to consume. By watching
these TV shows or consuming these cultural goods, she improved her CHL proficiency.
Similarly, Dianna and her family sometimes read Chinese books together at home. When
asked about her informal Chinese learning, she replied, “Sometimes at home, we have kind of
family discussion. We read a book together in Chinese and we discuss in Chinese.” She
considered this family reading and discussion as informal Chinese learning that may improve
her Chinese language. For Crystal and Dianna, the consumption of these Chinese cultural
goods can be understood as the investment of cultural capital, eventually convertible into
CHL proficiency.
In addition, all the participants interviewed indicated that their parents and family
members were very supportive of their CHL learning and encouraged them to learn Chinese
when they were small. These family commitments were valuable resources in these
participants’ CHL practice and could be understood as social capital that contributed to their
CHL proficiency.
Resonating with the findings from the quantitative phase, symbolic capital was also
beneficial for these participants in their Chinese practice. When asked about the driving force
behind his efforts in learning Chinese, Adam said:
The driving force, in a way, may be “jealousy”. I don’t know whether that’s the right
word to say. For example, my friends who are Malaysians or Singaporeans can speak in
Mandarin or in any other dialects, like Cantonese. I thought, “Hang on, if they can do it,
surely I can do it, right?”
279
This exemplifies a ‘face’ issue, as explained in Chapter Three. Adam attached symbolic value
to his ‘face’. This might be the reason why he felt jealous of his friends who could speak
Chinese. In this situation, it was a matter of ‘face’ that drove him to learn CHL.
In Bourdieu’s sense, the desire for symbolic capital was contributing to his CHL learning.
The initial quantitative phase and the subsequent qualitative phase revealed a two-way
relationship between capital and CHL proficiency. This conforms to the empirical findings in
the existing literature. As explained in Chapter Two, empirical studies have not explicitly
used the concept of capital to investigate CHL learning, but Bourdieu’s capital metaphor
would shed additional light on these empirical findings. The opportunities for CHLLs to be
exposed to CHL are determined by the quantity and quality of various resources available to
CHLLs (Lao, 2004; Luo & Wiseman, 2000; J. Zhang, 2009). The more opportunities to use
CHL in a variety of economic, cultural, and social contexts, the more vital CHL is. These
opportunities will shape present and future actions with regard to CHL learning. As a return
on this CHL learning, CHLLs may acquire wider access to more resources, e.g. more
employment opportunities and better career development (Francis et al., 2009; A. Hancock,
2006; Lao, 2004; Weger-Guntharp, 2006; Wong & Xiao, 2010); more exam credentials or
knowledge of China and Chinese culture (Francis et al., 2009; J. Zhang, 2009); better
communication with Chinese family members (Francis et al., 2009; Lao, 2004) and Chinese
friends (J. Zhang, 2009); deeper involvement in Chinese communities (Lao, 2004); as well as
the enhancement of reputation and the accumulation of awards (J. Zhang, 2009). In short,
CHLLs will capitalise on their CHL by investing various forms of capital. As a return on their
investment, their CHL, when valued and recognised as linguistic capital, can be converted
back into various forms of capital. It is through the two-way interaction between CHL
learning and capital (re)production that CHL proficiency realises its exchange value.
280
6.4.2 ‘Chineseness’ and CHL proficiency
Another resonance with the findings from the quantitative phase was the role played by
‘Chineseness’ qua habitus in CHL practice. As embodied dispositions, habitus is rooted in the
agents’ past. It underpins agents’ present perceptions and actions, and produces agents’
practices. Language choice as a form of practice can be quite unconscious sometimes. When
asked about the reasons for learning Chinese, Adam said, “I tried to learn Mandarin because I
just wanted to learn it”. He added, “I don’t know how it happened but it just happened. It’s a
good weird thing that had happened.” When asked the same question, Crystal said, “I like it
because I want to.” She paused then added, “I feel I just want to know it.” En-ning said, “I
don’t know what came over me but I decided I wanted to learn it in school.” For Adam,
Crystal, and En-ning, they did not have to know what drove them to learn CHL consciously.
Their body knew. When Bob was asked why he chose to learn Chinese in school, he said it
was “because of my Chinese heritage”. By this account, it was his Chinese heritage rooted in
his body that generated his intention for learning Chinese. As Bourdieu (1977b) argues,
agents’ actions are the product of habitus of which the agents may have no conscious mastery
because habitus always exceeds conscious intentions. As such, the habitus of ‘Chineseness’
produced these participants’ CHL practices, without either explicit reason or signifying intent,
to be none the less sensible and reasonable, and to be immediately intelligible and foreseeable,
and hence taken for granted. This ‘Chineseness’ was the immanent law laid down in these
participants by their earliest upbringing and “previous state” (Bourdieu, 2000, p. 161), which
were the preconditions for their CHL practices.
Being a constitutive dimension of ‘Chineseness’, looking Chinese is one of the bodily
attributes that belong to all participants interviewed. Chinese appearance of yellow skin and
black eyes and hair can be understood as contributing to the durable and transposable habitus
281
of ‘Chineseness’ that the participants have to carry with them all through their lives, willingly
or unwillingly. This visibility of looking Chinese is a biological fact for these participants, an
element of habitus that cannot be erased or made over (Luke, 2009). This biological essence
of ‘Chineseness’, to which the participants were involuntarily and inextricably attached,
connected them through their blood and flesh to their Chinese ethnicity, and linked them to
their CHL learning. As Dianna said, “I was born into this culture, this colour, and this
language.” Similarly, Crystal said, “We are Chinese family and we should speak Chinese.”
Dianna had an experience deeply inscribed in her memory that her friends in high school told
her: “If you are Asian, you should be able to speak your home language. If you can’t, it’s sort
of shame.” This point was shared by En-ning. She said, “How difficult it is for a lot of
overseas Chinese and how shameful it is! They look Chinese but they don’t speak Chinese.”
Adam explained his similar experience:
As far as I know, all my grandparents are Chinese. Of course, I will look like Chinese
anyway. If I try to speak Mandarin, they don’t say, “Wow, you can speak Mandarin.”
They tend to go, “How come you can’t speak Mandarin?” …So I guess people around
me tend to say, “Oh, if you look like a Chinese, you can speak Mandarin.”
He also recalled his experience when he lived in Singapore:
Even in Singapore, the taxi drivers sometimes looked at me and said …what are the
words…I tried to copy what they said in Mandarin…“你是华人么 (Are you Chinese)?”
If I said yes, they said “你说你是华人,为什么你不可以讲华语 (You said you are
Chinese, why you can’t speak Chinese)?”
282
For Adam, these experiences constructed the assumption that looking Chinese meant being
able to speak Chinese. It was this stereotypical perception deeply rooted in his mind that
largely drove his conscious commitment to Chinese learning. He tried to learn Mandarin in
Year 12 but the school did not offer it. He was very keen to learn Chinese so he decided to
learn Japanese Kanji in school, which was “the closest thing” to Chinese. As he explained,
“They have the Kanji which is supposed to be similar to 汉字 (Chinese character). So I
thought that’s ok. I will just do it.” He reported that Japanese Kanji became the first Chinese
characters that he was exposed to. In daily life, he tended to “train” and “force” himself to
read Chinese news and watch Chinese TV, and tried to “expose” himself “as much as possible”
to Chinese. For Adam, it might not be the return on the investment of various forms of capital
but his embodied dispositions of ‘Chineseness’ that drove him to pursue CHL. This was
reflected in the statistical models in the initial quantitative phase whereby his CHL
proficiency was predicted by his habitus of ‘Chineseness’ instead of various forms of capital
that he invested in CHL practice. En-ning also considered this habitus of ‘Chineseness’ an
internal driving force behind learning CHL. She reported, “The difference is between a white
person learning Chinese and the overseas Chinese learning Chinese.” When asked to explain
this, En-ning added, “For Chinese Australians learning Chinese, learning the language
immediately raises issues of identity, belonging, culture, and history. I think this means they
are often highly motivated because of the numerous factors driving them to learn.” For
En-ning, ‘Chineseness’ seemed to be associated with Chinese ethnic identity, a sense of
belonging to Chinese cultural history and heritage. She considered these dimensions within
‘Chineseness’ to be the factors driving Chinese Australians to learn CHL.
In brief, these participants’ ‘Chineseness’ as their habitus operated at a level that was
simultaneously conscious and unconscious. They did make language choices in strategic
283
ways, and try to use the rules of different fields to their advantage, but at the same time they
were influenced, or almost driven, by values and expectations that they derived from their
habitus. Though they might be conscious of learning Chinese as their HL strategically, they
might not be aware that their motives, goals, and aspirations were generated through their
habitus of ‘Chineseness’.
6.4.3 Quantitative coda
In this thesis, the findings from the initial quantitative phase and the subsequent qualitative
phase have supported the claim of the mutually constitutive effects between ‘Chineseness’
and CHL proficiency, and between ‘capital’ and CHL proficiency. These findings conformed
to Bourdieu’s theorisation that habitus and capital make sense of people’s language choices
and practices in particular fields. These findings also prompted the question of whether
‘Chineseness’ and ‘capital’ were related to other social practices of the participants, such as
their language usage at home, years of formal CHL learning, frequency of visits to China, and
self-perceptions of who they were. These issues emerging from both the quantitative phase
and the qualitative phase are discussed below.
Spearman’s Correlation Test and Kendall’s Correlation Test were conducted. ‘Years of
formal CHL learning’ and ‘frequency of visits to China’ were variables measured at interval
level. Correlation between ‘Chineseness’ and these two variables were examined by
Spearman’s Correlation Test. Language usage at home and self-labelling were variables
measured at ordinal level, ranked in three levels and five levels respectively. Consequently,
there were many scores with the same ranking. As such, correlation between ‘Chineseness’
and these two variables were examined by Kendall’s Correlation Test. Results demonstrated
that ‘Chineseness’ had a statistically significant small to medium positive correlation with
284
language usage at home (τ = .44, p < .001), years of formal CHL learning (r = .43, p < .001),
frequency of visits to China (r = .48, p < .001), and self-labelling (τ = .43, p < .001). Likewise,
the relationship between different forms of capital and these social practices was also
examined. The results demonstrated that different forms of capital had a statistically
significant small to medium positive correlation with these social practices. The correlation
matrix is reported in Table 6.3.
Table 6.3
Correlation between forms of capital and social practices
Language usage
at home
Years of formal
CHL learning
Frequency of
visits to China Self-labelling
Economic capital τ = .31, p < .001 r = .44, p < .001 r = .51, p < .001 τ = .30, p < .001
Cultural capital τ = .39, p < .001 r = .46, p < .001 r = .47, p < .001 τ = .39, p < .001
Social capital τ = .43, p < .001 r = .46, p < .001 r = .48, p < .001 τ = .43, p < .001
Thus, the stronger participants’ ‘Chineseness’ was and the more value of capital participants
accrued, the more likely they were to use Chinese at home, to spend more years on formally
learning Chinese, to visit China more frequently, and to consider themselves more Chinese
than Australian. Likewise, when participants were more engaged in these social practices,
their habitus of ‘Chineseness’ was more likely to be reinforced or recaptured, and more
capital was to be produced or reproduced. In brief, there is a mutual relationship between
‘Chineseness’/capital and social practices associated with CHL learning.
It is important to highlight the correlation between ‘Chineseness’ and the frequency of
visits to China. In this respect, participants’ visits to China were linked to their ‘Chineseness’.
On the one hand, these physical returns, to some degree, were generated by their
‘Chineseness’. On the other hand, the physical returns have become a crucial process in
building and shaping their ‘Chineseness’. As such, China was not a place simply left behind,
but a concurrent and ongoing social space of cultural and historical attachment. The
285
qualitative data also resonated with this finding. During his Hong Kong trip, Bob tended to be
interested in his family roots. He said:
So maybe it’s just some specific parts that I found quite interesting to me. I found my
great grandmother was from a people called Hakka in Hong Kong, who were foreigners
but came to Hong Kong back in the early days. Maybe if I looked into that, it’s quite
interesting.
The Chinese characters for Hakka are 客家, which literally mean ‘guest families’. As the
name indicates, the Hakka people have a tradition of migration, and hence were considered
‘guests’ in the places where they have moved to. Hakka communities were thought to have
arrived in Hong Kong in the late 17th
century (Hase, 1995). Since they were not the original
residents of Hong Kong, Bob used the word “foreigners” to describe them. Although Bob
claimed that the two-week Hong Kong trip was not as exciting as he had expected, he thought
the discovery of his great grandmother’s origin was “quite interesting”. It will be recalled that
En-ning also reported her “incredible” trip to her grandparents’ hometown in China. These
data indicated that Bob and En-ning considered the discovery of the roots of their cultural
history meaningful for their present sense of self and belonging. As indicated in previous
research (Carreira, 2004), cultural heritage and historical background nurture and reinforce
ethnic identity.
It will be recalled that all the interviewees reported that their parents insisted that they
learn CHL when they were small and likewise, they themselves would encourage their next
generation to learn CHL as a way of maintaining the habitus of ‘Chineseness’. By this
inculcation, the habitus of ‘Chineseness’ becomes durable and transposable from generation
286
to generation. However, habitus is not immutable. This led to the investigation of
generational difference in habitus. Thus, the relationship between respondents’ ‘Chineseness’
and their generation was examined.
Two groups were firstly identified, namely 121 Australian-born participants (first
generation) and 109 non-Australian-born participants (later generation). Before comparing
these two groups, the statistical assumptions were checked. Normality was first checked by
Kolmogorov-Smirnov Test. The result demonstrated a non-significant (p = .200)
Kolmogorov-Smirnov score of .06 for both groups. The assumption of normality was
satisfied. The Levene’s Test for Equality of Variances demonstrated a non-significant (p
= .119) F value of 2.45, which satisfied the assumption of homogeneity of variance.
Satisfaction of these assumptions suggested the basis for t-test. Levels of ‘Chineseness’ of the
non-Australian-born group were shown to be on average higher than those of the
Australian-born group (t = 5.24, p < .001). This result presented a medium-sized effect (r
= .33). This genuine effect existing in relation to different birthplaces of Chinese Australians,
or social space in Bourdieu’s term, will be discussed in the final chapter.
Secondly, the Australian-born participants were further divided into second-, third-,
and fourth-generation and further removed. When generation was treated as a variable
measured at ordinal level, ranked in four levels, Kendall’s Correlation Test indicated a
statistically significant small to medium negative correlation between ‘Chineseness’ and
generation (τ = -.38, p < .001). In other words, ‘Chineseness’ of later generations was weaker
than that of earlier generations. These findings indicated that participants’ habitus of
‘Chineseness’ was not immutable. The habitus of ‘Chineseness’ associated with different
generations of Chinese Australians was shaped in relation to the passage of time. This will be
discussed in Chapter Seven.
287
6.5 Chapter summary
This chapter rendered a comprehensive analysis and interpretation of the data emerging from
the interviews. Findings relating to the meanings that participants attached to CHL were
interpreted through the theoretical framework. By their accounts, CHL competence can
produce new capital for all participants as a return on their investment. Firstly, all participants
indicated that CHL competence could create more and better job opportunities in future
labour markets. As such, their CHL competence has the potential to produce economic capital.
Secondly, participants’ CHL competence contributed to their capture of different forms of
cultural capital. Objectified cultural capital was produced through their CHL learning,
existing in the forms of purchased cultural goods, such as Chinese textbooks and books, as
well as Chinese TV shows. Embodied cultural capital, such as gaining Chinese cultural
knowledge, being able to sing Chinese songs, and being adapted to Chinese ways of doing
and thinking, was captured through participants’ CHL learning. Institutionalised cultural
capital, e.g. educational credits, was also produced through participants’ CHL learning.
Thirdly, CHL competence helped the participants build social capital in Chinese-speaking
worlds, including better communication with family members, more contacts within
friendship networks, and more engagement in social lives. Fourthly, CHL competence
benefited the participants in terms of access to recognised awards, honour, and pride, all of
which can be understood as forms of symbolic capital. In addition to the production of new
capital, CHL competence enabled the participants to claim their ‘Chineseness’. As such, CHL
competence was associated with participants’ habitus. In conclusion, CHL practice offered
meaningful benefits for these participants in the production or reproduction of various forms
of capital and the reinforcing or recapture of their habitus of ‘Chineseness’. Table 6.4
summarises the main findings from the interview data.
288
Tab
le 6.4
Interv
iew d
ata
sum
mary
Field
Hab
itus
Sym
bolic
capital
Social
capital
Cu
ltural
capital
Eco
nom
ic
capital
Con
cepts
Caterin
g to
the field
of fo
rces
Differen
t valu
es in d
ifferent situ
ations
Co
nscio
us actio
ns
Unco
nscio
us d
ispositio
ns
Em
bodied
disp
ositio
ns
Fam
ily ro
ots
Aw
ards
Reco
gnitio
n b
y p
eople
Socialisatio
n w
ith p
eople
Co
mm
unicatio
n w
ith frien
ds
Co
mm
unicatio
n w
ith fam
ily m
embers
Form
al edu
cation cred
its
Ch
inese w
ay o
f doin
g th
ings
Und
erstandin
g o
f Chin
ese cultu
re
Ch
inese cu
ltural g
oods
Job
oppo
rtunities
Ben
efits / Mea
nin
gs o
f CH
L
Th
e forces o
f fields th
at structu
re agen
cy an
d sh
ape ag
ents’
practices in
line w
ith th
e rules o
f fields
A so
cial place co
ntain
ing ag
ents stru
gglin
g fo
r po
sitions b
y
the co
ntro
l of reso
urces
Imm
anen
t pro
pen
sities linked
to cu
ltural h
istory
that d
rive
the actio
ns b
oth
inten
tionally
and u
nin
tentio
nally
Rep
utatio
ns fo
r com
peten
ce and
imag
es of resp
ectability
and
honourab
ility
Co
ntacts
and
gro
up
mem
bersh
ips
that
pro
vid
e actu
al or
po
tential su
pport an
d access to
valu
ed reso
urces
Institu
tionalised
cultu
ral capital: C
ultu
ral com
peten
ce with
a
conven
tional,
constan
t, leg
ally
guaran
teed
valu
e in
connectio
n to
certain in
stitutio
ns
Em
bodied
cultu
ral capital: L
asting
disp
ositio
ns o
f min
d an
d
body, ex
isting in
the fo
rm o
f schem
ata of p
erceptio
ns an
d
action
s
Ob
jectified cu
ltural cap
ital: Material fo
rms o
f cultu
ral goods
Co
nvertib
le into
mon
ey
Cod
ing criteria
√
√
√
√
√ √
√ √
√
√
√
Ad
am
√
√
√ √
√ √
√ √
√ √ √
Bo
b
√
√
√
√ √
√
√
√
√
√
√
√
√
Cry
stal
√
√
√ √ √
√
√
√
√ √
√
Dia
nn
a
√
√
√
√
√
√
√
√
√
√
√
√ √
En
-nin
g
289
Furthermore, the findings from the interview data conformed to the findings from the
initial quantitative phase. Data of the initial online questionnaire have revealed that both
participants’ habitus of ‘Chineseness’ and their invested capital can contribute to their CHL
proficiency. Similar findings emerged from the subsequent interview data. As summarised in
Table 6.5, money spent on CHL learning, parents’ and friends’ supports in CHL learning,
consumption of Chinese cultural goods in the process of CHL learning, face issues in CHL
learning, and conscious and unconscious dispositions of ‘Chineseness’ all contributed to CHL
proficiency. In Bourdieu’s sense, habitus and various forms of capital made sense of the
choice of CHL in given fields.
Table 6.5
Contributing factors to CHL proficiency
Concepts Contributing factors Adam Bob Crystal Dianna En-ning
Economic
capital
Money spent on CHL learning,
e.g. textbooks, tuition, visit to
China, and Chinese TV shows
√ √ √
Cultural
capital
Consumption of Chinese cultural
goods through CHL learning, e.g.
watching Chinese TV shows and
reading Chinese books
√ √
Social
capital
Family and friends’ supports in
CHL learning √ √ √ √ √
Symbolic
capital Recognition by people √
Habitus
Embodied dispositions √ √ √ √ √
Unconscious dispositions √ √ √
Conscious actions √ √ √ √ √
In addition, the findings led to new questions: Were participants’ ‘Chineseness’ and
‘capital’ related to a wider range of their practices? Was there a generational difference in
participants’ ‘Chineseness’? Further examination of the quantitative data offered answers to
these questions. This quantitative coda revealed that participants’ habitus of ‘Chineseness’
and their ‘capital’ were correlated to a wider range of their practices, such as language usage
at home, years of formal CHL learning, frequency of visits to China, and their
290
self-perceptions of who they were. The quantitative coda also discovered the generational
difference in participants’ ‘Chineseness’, with later generations having weaker levels of
‘Chineseness’.
In summary, participants’ CHL competence produced through various social practices
associated with CHL learning at given times across certain places was recognised as
linguistic capital with different value in different fields. Production or reproduction of various
forms of capital and reinforcement or recapture of the habitus of ‘Chineseness’ happened in a
process of exchange, an exchange of value and forms of resources. Participants’ various
forms of capital and their habitus of ‘Chineseness’ were convertible into their CHL
competence and reconvertible into themselves in identical or expanded forms within the field
of exchange. It was through this process that participants’ CHL competence, as linguistic
capital, realised its exchange value.
291
Chapter 7: Discussion and Conclusion
This chapter offers the discussion and conclusion of the research reported in this thesis.
Section 7.1 discusses the findings from the quantitative phase and the qualitative phase;
explains these findings by virtue of Bourdieu’s theoretical triad of habitus, capital, and field;
and talks back to the research questions. This section also discusses several important issues
emerging from the findings. These issues include: how habitus and capital related to other
social practices besides CHL learning; how generation impacted on ‘Chineseness’ across time
and space; how bodily emotion was generated through power relations; and how symbolic
power and symbolic violence were imposed within given fields. Section 7.2 highlights the
theoretical, methodological, and practical contributions of this study. Section 7.3 outlines the
limitations of the study. Section 7.4 suggests directions for future research. The last section
concludes the thesis with a succinct summary. It recalls the research journey over the life of
the thesis, talks back to the existing literature, and revisits the researcher’s subjectivity.
7.1 Discussions
Bourdieu’s (1989a, p. 101) equation “[(habitus) (capital)] + field = practice” summarises the
conceptually and empirically essential relationship amongst capital, habitus, and field, which
illuminates individuals’ practices. Specifically, Bourdieu (1991) theorises that people make
language choices as a form of social practice according to the amount of different resources
(capital) as well as the dispositions (habitus) they have within a given field. This theorisation
is the underpinning basis for the theoretical framework and the research design of this study.
Guided by Bourdieu’s sociology, the study investigated the entanglement amongst the habitus
292
of ‘Chineseness’ of young Chinese Australian adults, this group’s investment of capital in
CHL learning, and their CHL proficiency. Several important issues emerged from this
investigation. These issues are discussed below.
7.1.1 Habitus, capital, and practice in fields
In Chapter Five, SEM tested and explained that both ‘Chineseness’ and ‘capital’ made
statistically significant contributions to ‘CHL proficiency’. In the multiple regression model,
multicollinearity between ‘Chineseness’ and ‘capital’ was diagnosed from a statistical
perspective. This problem can be further illuminated from a theoretical perspective. As an
embodied property and “incorporated and quasi-postural disposition” (Bourdieu, 1985a, p.
13), habitus serves as a form of “capital” (Bourdieu, 1985a, p. 13), specifically a cultural
capital (Bourdieu, 1973). Accordingly, the habitus of “‘Chineseness’ becomes a form of
cultural capital” in particular contexts (A. Louie, 2004, p. 21; Ong, 1999, p. 6). In order to be
a form of capital, one’s habitus of ‘Chineseness’ must be valued and recognised within a
particular field. ‘Chineseness’ as a set of embodied dispositions associated with certain
Confucian values can be understood as cultural capital when these embodied dispositions of
‘Chineseness’ are accredited by the legitimate Confucian cultural tastes and inclinations.
Specifically, ‘Chineseness’, when valued, can be understood as embodied cultural capital
because it has been integrated within the lasting dispositions of mind and body, existing as
perceptions and behaviours of the participants. In this respect, ‘Chineseness’ was embedded
in the dimensions of cultural capital. Young Chinese Australian adults may develop their
‘Chineseness’ in relation to how much embodied cultural capital they have in a Chinese
cultural or social field. This relationship is very important because it reveals how the amount
of capital that they have can condition their dispositions of being, doing, and thinking.
293
When taken account of together, participants’ habitus of ‘Chineseness’ and their
various forms of capital became the overarching drivers behind the CHL practices that
produced CHL proficiency. Moreover, participants’ ‘Chineseness’ and capital produced other
social practices, such as language usage at home, years of formal CHL learning, frequency of
visits to China, and self-labelling, which in return reinforced ‘Chineseness’ and generated
various forms of capital. Data analysis in the quantitative coda demonstrated that language
usage at home, years of formal CHL learning, frequency of visits to China, and self-labelling
had a statistically significant small to medium positive correlation with ‘Chineseness’ and
capital. However, this ‘Chineseness’ and capital did not generate the same practices for every
Chinese Australian. Data analysis in the quantitative coda demonstrated a wide range of
variance in language usage at home (from English only to Chinese only, with different levels
of mixed usage in between); years of formal CHL learning (from none to over 15 years);
frequency of visiting China (from none to over 20 times); and self-labelling (from
considering themselves basically Australian to basically Chinese, with degrees of a blend of
both in between). As such, depending upon the stimuli and structure of the field, habitus and
capital will generate, but will not determine, outcomes (Bourdieu & Wacquant, 1992).
Interestingly, the social practices had the strongest correlation with the form of capital
to which they had most relevance. Language usage at home had the strongest correlation with
social capital in the Chinese community (r = .43, p < .001). When family connection became
a valued social capital, participants tended to use more CHL to maintain and strengthen their
family ties. Self-labelling also had the strongest correlation with social capital in the Chinese
community (r = .43, p < .001). When involvement in the Chinese community accrued more
value, participants tended to have stronger feelings of belonging to this community.
Consequently, they considered themselves more Chinese. Frequency of visits to China had
294
the strongest correlation with economic capital (r = .51, p < .001) because international travel
cost participants money. It also had a small to medium correlation with cultural capital (r
= .47, p < .001) and social capital (r = .48, p < .001) because participants were culturally and
socially connected to China, a place of origin, roots, and heritage. In brief, there is a two-way
relationship between ‘capital’ and participants’ various social practices associated with their
CHL learning. Participants invested various forms of capital in their CHL learning through
these social practices. As a return on their investment, their CHL proficiency improved.
When their CHL proficiency was recognised and legitimised as linguistic capital in certain
fields, this linguistic capital realised its exchange value, accrued added value, and produced
various forms of capital.
Of all the correlations discussed above, it is particularly important to highlight the
relationship between ‘Chineseness’ and visits to China, which fostered and maintained
participants’ ties to their original roots and played a crucial and ongoing role in informing
their notions of ‘Chineseness’. This finding echoed A. Louie’s (2004) study in which her
Chinese American participants reportedly “reinforced” (p. 110) their ‘Chineseness’ through
visits to their ancestral villages in China and one participant reported, “As soon as my feet
touched China, I became Chinese.” (p. 116) Literally, homeland can be interpreted as a
concrete place where intimate familial and social relations are established through the lived
experience of locality and community. For Chinese Australians, China is more than a
geographical place away from their home in Australia. It is also a homeland infused with
meanings and relevance that create a sense of ancestral origin, historical roots, and cultural
heritage. Traditionally, migration was a one-off process of movement from homeland to
hostland. At present, it is becoming easier than ever before for Chinese Australians to travel
back and forth between China and Australia. With regard to space, these movements function
295
as a form of deterritorialisation that enables co-existing homes in China and Australia. With
regard to time, such physical mobility allows Chinese Australians to continually move
between their past and their present.
In relation to space, the non-Australian-born group had emigrated from China or from
countries with Chinese population. They had spent their early years in social spaces with
greater access or engagement in Chinese language and culture than the Australian-born group,
suggesting a deeper understanding of and stronger sense of attachment to Chinese culture
than the Australian-born group. As Bourdieu (1991) acknowledged, cultural dispositions may
be acquired by prolonged exposure to a culture, or through the deliberate inculcation of a
culture. In line with this Bourdieusian perspective, ‘Chineseness’ can be understood as a set
of culturally learned dispositions acquired through experiences of daily life and structured
through activities within social fields that featured Chinese culture. The dispositions
associated with the habitus of ‘Chineseness’ of the non-Australian-born group were
inculcated by their Chinese cultural history prior to their immigration and then transposed
across different social spaces, from their birthplace to Australia. This could explain their
stronger expressions of ‘Chineseness’ than the Australian-born group (t = 5.24, p < .001, r
= .33).
In relation to time, the ‘Chineseness’ of later generations was significantly weaker than
that of more recent generations. The family of participants from later generations had longer
settlement periods in Australia than that of participants of more recent generations. Later
generations did not have as many Chinese cultural memories and experiences as the more
recent generations. Their physical and cultural ties to their Chinese roots were less intense.
As argued earlier, habitus is lasting but not immutable. With the passage of time, different
generations made gradual transitions from being migrants to becoming integrated members of
296
the host society as their lives gradually took root in a land away from the cultural home. They
tended to internalise the values and imperatives of the host society. Consequently, a habitus
of ‘Chineseness’ can be expected to gradually fade generation by generation (τ = -.38, p
< .001). This resonates with Luke’s (2009) position that hyphenated and diasporic identity
necessarily leads to imperfect intergenerational reproduction of embodied dispositions.
With the change of life worlds and the passage of time, different generations
experience different culturally and historically situated upbringings, which in turn can shape
a different habitus. Habitus is thus at the basis of strategies that tend to maintain these
differences, hence concurring in different practices (although not consciously or deliberately)
in reproducing the entire system of differences across time and space (Bourdieu, 1996). The
difference in participants’ ‘Chineseness’ thus led to different CHL practices, in terms of
different levels of effort put into learning CHL, formally or informally, willingly or
unwillingly, and consciously or unconsciously. Consequently, the difference in ‘Chineseness’
could be understood to produce the difference in CHL proficiency.
These findings echoed previous studies. First-generation CHLLs were more proficient
in CHL than their counterparts from later generations (Kiang, 2008). There was a significant
association between generation and feelings of being Chinese, as well as between
self-labelling and ethnic exploration (Rosenthal & Feldman, 1992). Over time and across
space, there was erosion of ethnic identification, behaviours, and knowledge but not of the
importance and evaluative components of ethnic dispositions (Rosenthal & Feldman, 1992).
In general, later generations do not maintain ‘Chineseness’ to the same level of
intensity as more recent generations because the habitus of ‘Chineseness’ is not immutable.
However, some sense of ‘Chineseness’ does persist within the Chinese ethnic group because
the intergenerational influences create a sense of maintaining ‘Chineseness’ that makes the
297
habitus of ‘Chineseness’ durable and transposable. In this respect, ‘Chineseness’ describes
the stronger or weaker allegiance of Chinese Australians to their ethnic group, with which
they have ancestral links. Although there is no necessity for continuation over time and
across space of exactly the same socialisation or cultural patterns, it is in a relation of
diversity within homogeneity that the shared habitus of the different members of the same
group are unified (Bourdieu, 1977b).
7.1.2 Field of forces
As argued in Chapter Three, a field is a structured social space, “a field of forces” (Bourdieu,
2011, p. 40). The forces within a field structure the power relations between agents who
dominate and those who are subordinate. The interview participants were conscious of their
subordinate or dominated positions in particular situations. Consequently, their feeling of
‘shame’, generated by experience of subordination, was repeatedly emergent from the
interview data. This is consonant with previous research in which participants considered not
being able to speak Chinese as a “disgrace”, “embarrassment”, and a reason for being
“ashamed” (Francis et al., 2009, p. 529). From a Bourdieusian perspective, this ‘disgrace’,
‘shame’, and ‘embarrassment’ can be understood as forms of bodily emotion, “the practical
recognition through which the dominated, often unwittingly, contribute to their own
domination by tacitly accepting, in advance, the limits imposed on them”, often taking the
forms of shame, timidity, anxiety, guilt, blushing, inarticulacy, clumsiness, and trembling
(Bourdieu, 2000, p. 169). These forms are ways of reluctantly submitting to the dominant
judgement, “sometimes in internal conflict and ‘self-division’, the subterranean complicity
that a body slipping away from the directives of conscious and will maintains with the
violence of the censures inherent in the social structures” (Bourdieu, 2000, pp. 169-170). It
298
will be recalled that in Chapter Six, En-ning reportedly felt ashamed for her “funny accent” in
a situation where Mandarin was the dominant and legitimate language. The lack of quantity
and quality of Mandarin, or linguistic capital in Bourdieu’s term, placed En-ning in a
subordinate position. En-ning, therefore, considered not being able to speak Chinese
“shameful”. Likewise, Adam apologised for his tones during the interview. In this case,
subordination emerged from hierarchies based on mastery of spoken Chinese. Adam gave
evidence of a sense of subordination through a possibly unconscious apology for his lower
Chinese proficiency, compared to the researcher’s native Mandarin proficiency. In these
examples, there were indeed moments where participants demonstrated the feeling of ‘shame’,
the bodily emotion of anticipating an experience of being dominated in a given field. In
contrast to Adam and En-ning’s experiences, Dianna was able to slip away from being
subordinate or dominated when her Chinese proficiency was recognised. In response to this
recognition, she did not feel “any shame at all” for being able to speak Chinese.
In relation to the forces within fields, participants also reported that their parents forced
them to learn CHL when they were small, their Chinese teachers set a higher standard for
them than their white peers, and people around them held the stereotype that looking Chinese
meant being able to speak Chinese. This is consistent with previous research that indicates
people with more power can construct “practical taxonomies of ways of seeing and doing in
the world” (Lane, 2000, p. 195). Parents’ insistence, teachers’ expectations, and people’s
assumptions were norms and conventions that granted parents, teachers, and those people
legitimate power and made them more powerful others to set the rules within particular fields.
These rules were believed, obeyed, and respected by the participants. Consequently, parents,
teachers, and people accrued symbolic capital, the possession of which constructs the
foundation of symbolic power (Bourdieu, 1989b). Parents, teachers, and people could wield
299
their symbolic power to ‘force’ their children to learn CHL, to set a higher standard for
CHLLs, and to expect Chinese-looking people to speak Chinese. Symbolic power is the
power to consecrate and reveal things with words (Bourdieu, 1989b). As such, parents,
teachers, and people may exert their symbolic power by persuasion, inculcation, and
discourse to legitimise CHL as a valuable resource in a particular social field. At the same
time, participants granted parents, teachers, and people the legitimacy and right to impose
their expectations. Bourdieu (1977a, p. 648) explained this power relation, “Those who speak
regard those who listen as worthy to listen and those who listen regard those who speak as
worthy to speak”.
This power relation can be understood as constituted within and by the rules and forces
within fields. As agents, the participants had practices that were structured by these forces of
the fields. Due to their agency, they tended to adjust to the structure of the field and submit to
these forces. Since these forces were imposed on them, their participation was quite often
reluctant. In this respect, these forces can be understood as symbolic violence.
The notion of symbolic violence is particularly relevant to the Chinese cultural and
social context. As discussed in Chapter Three, Confucianism has consecrated and legitimised
the power of parents and teachers by its core values of Zhong (忠, Loyalty), Xiao (孝, Filial
Piety), and Li (礼, Ritual or Propriety). These Confucian values position parents and teachers
at the top of the social hierarchy. In this respect, symbolic violence describes the tacit and
explicit modes of cultural and social domination by parents and teachers occurring within the
social lives of participants. Specifically, it accounts for this imposition of parents’ and
teachers’ will to urge CHL learning upon the less powerful participants, and the submission
to this imposition, willingly or unwillingly, by the participants. The symbolic violence that
imposes CHL learning on CHLLs was also revealed by existing literature: Parents forced
300
their children to learn CHL when their children were small (Weger-Guntharp, 2006; Wong &
Xiao, 2010); Chinese appearance was taken for granted as a sign of being able to speak
Chinese (Ang, 2001); some early generations of American Cantonese (A. Louie, 2004) and
British Cantonese (Watson, 1975) have the common sentiment that one could not be
considered truly Chinese without receiving an education in Chinese; and some Chinese
Canadians learn CHL because of various external “pressures” (Comanaru & Noels, 2009, p.
148).
Empirically, participants might consider themselves free individuals yet sometimes had
to base their everyday practices on the assumptions about dispositions, behaviour, and
attitudes of powerful others. Sociologically, their social practices (agency) could be explained
by social regularities (structure). These both raise fundamental questions that Bourdieu’s
notions of symbolic power and symbolic violence help to resolve and explain the explicit
rules dictating such practices. On the one hand, symbolic power and symbolic violence were
structures within fields that pushed participants to learn CHL. On the other hand, the external
symbolic power and symbolic violence were gradually internalised by the participants’
agency as their own habitus and capital to continue to pursue CHL as a linguistic capital.
7.2 Contributions of the study
There has been a recent surge of research interest in HLLs (Lynch, 2003). However,
compared to other HLLs, a relatively small body of HL research addresses issues particular to
CHLLs (Levesque, 2007; D. Li & Duff, 2008). Considering the size, vitality, and growth of
the CHL learning population, research with CHLLs warrants more attention (He, 2008; D. Li
& Duff, 2008). This mixed methods study, guided by Bourdieu’s theory of habitus, capital,
and field, has offered theoretical, methodological, and practical contributions.
301
7.3.1 Originality and contributions to new knowledge
HL research is an emerging field that has been gaining increasing attention (Lo-Philip, 2010).
The reciprocal and mutually constitutive effects between HL and ethnic identity have been
well documented in the literature (Cho, 2000; Cho et al., 1997; Fishman, 1977; Giles et al.,
1977; Giles & Johnson, 1987; Hurtado & Gurin, 1995; Smolicz, 1981), with studies
providing evidence about Chinese Americans/Canadians, Korean Americans/Canadians,
Japanese Americans/Canadians, Armenian-, Latin-, Asian-, Mexican-, and
Vietnamese-Americans, and Italian-, Portuguese-, Ukrainian-, and Jewish-Canadians.
Positive correlation between learners’ ethnic identity and their HL has been found by a large
number of social psychological studies in North America. However, there is scant
sociological research conducted outside North America on such topics. This study has
contributed theoretically to the literature by conceptualising and operationalising Chinese
Australians’ ‘Chineseness’ as their ethnic identity and constructing the essential link between
their ‘Chineseness’ and their CHL proficiency through Bourdieu’s sociological notion of
habitus.
There is an extensive tradition of studies on commitment to Second Language learning,
understood in psychological term as ‘motivation’ (Gardner, 1968, 1985; Gardner & Lambert,
1959, 1972). In contrast, the body of sociological work investigating commitment to Second
Language learning in general is smaller and more recent. As a sociological alternative to the
psychological concept of ‘motivation’, the concept of ‘investment’ (Norton, 1995) in Second
Language learning has offered a theoretical breakthrough (Canagarajah, 2006; Pavlenko,
2002). ‘Investment’ demands ‘capital’ (Bourdieu, 1986). To capture the complex process of
language learning and to offer an adequate investigation into learners’ trajectories, the notion
of capital has found wide purchase in the language education world. Following this route,
302
another theoretical contribution of this study was the use of Bourdieu’s sociological notion of
capital in a HL context to interpret Chinese Australians’ investment of available resources in
their CHL practice.
Habitus, capital, and field are the three fundamental thinking tools of Bourdieu’s
sociological approach. This study realised the full value of a Bourdieusian perspective by
constructing an inter-dependent and co-constructed triad of habitus, capital, and field to
examine language practice in a HL context. Moreover, this study extended Bourdieu’s
theorisation of language choice within a single language in a monolingual setting to language
choice among different languages in a multilingual context. A comprehensive theoretical
framework that can handle the interdependence of Chinese Australians’ ethnic identity, their
commitment to CHL learning, and their perceptions of CHL proficiency has been proposed.
7.3.2 Methodological contributions
In addition to the above theoretical contributions, the study also offered significant
methodological contributions. Much of the existing work attempting to apply Bourdieu’s
notions of capital, habitus, and field to people’s identities and commitment to language
learning has tended to be qualitative in approach (Connolly, 2011). However, there is
inevitably something missing by restricting the methodological focus to one mode of research
when operating Bourdieu’s theoretical package. In Bourdieu’s original works, such as
Distinction (Bourdieu, 1989a) and The State of Nobility (Bourdieu, 1996), data were also
quantitatively investigated by correspondence analysis. Though these data were mostly
analysed by a descriptive and exploratory multidimensional scaling technique, Bourdieu was
interested in tendency, prediction, and correlation based on these quantitative analysis. He
suggested (1990) that quantitative expressions of the distribution of capital in its different
303
forms among individuals can be developed and thus capital should be measurable. Likewise,
if habitus does represent the internalisation of broader social structures and does represent a
set of dispositions that are manifest, to some extent, in particular ways of thinking and
behaving, these wider patterns should be discernible and ultimately measurable at least to a
certain degree. Bourdieu’s concepts represent a methodological challenge in quantifying, but
given the increasing use of these concepts in sociology, such a methodology is needed
(Cockerham & Hinote, 2009). Following this route, a set of instruments was developed and
validated to quantify Bourdieu’s key concepts of capital and habitus within certain social
fields, and to investigate the research problems in this study.
Due to the technical limitations of the times, Bourdieu was not able to analyse the
quantitative data by the statistical models that have been developed today. The current study
has the benefit of statistical software and new modelling techniques to advance such lines of
enquiry. Moreover, Bourdieu’s key concepts can be investigated in a deeper manner by the
application of mixed methods design. Connolly (2011), informed by the work of Bourdieu,
suggests the need for greater use of quantitative methods in conjunction with in-depth
qualitative methods to further the understanding of the influence of identity in people’s lives.
In this respect, the current study made a methodological contribution to introduce a mixed
methods design that applies a Bourdieusian stance to the current research context.
7.3.3 Practical contributions
The pioneering research on HL dates back more than half a century, when Fishman (1964)
inquired into the field of minority language maintenance and shift. There were few echoes of
Fishman’s (1964) work until very recently when minority peoples’ HL and their ethnic
identities were recognised not only as assets to the individual minority people but also as
304
valuable resources to the social world around them (Brecht & Ingold, 2002; Campbell &
Rosenthal, 2000; He, 2006; Lynch, 2003; Peyton et al., 2001). Existing research argues
convincingly that developing HL speakers’ linguistic and cultural knowledge to advanced
levels is valuable not only for the HLLs themselves, their families, and communities, but also
for society more broadly (Fishman, 1991).
Firstly, understanding Chinese Australians’ heterogeneity and the potential drivers
behind CHL learning helps to better accommodate linguistically diverse learners in language
courses. As noted in Chapter One, there are a number of Confucius Institutes and Confucius
Classrooms established in Australia. There is now at least one Confucius Institute in all state
capitals of Australia. Confucius Classrooms in primary and secondary schools are also
proliferating. The courses offered by these Confucius Institutes/Classrooms are designed to
meet the demands of Foreign Language Learners of Chinese. However, many current,
prospective, and potential Chinese learners in Confucius Institutes/Classrooms are in fact
Chinese Australians who are CHLLs, as distinct from Foreign Language Learners of Chinese.
This highlights the importance of flexible teaching methods and pathways that cater to
Chinese learners’ varied learning backgrounds (Australian Government, 2012). By examining
Chinese Australians’ identities and their commitment to CHL learning, the current research
will be useful to Hanban and Australian institutions in partnership with Hanban, with an
interest in teaching and learning Chinese not only as a Foreign Language but also as a HL in
Australia. The research will help to develop Chinese language teacher training, Chinese
textbook and curriculum design, and Chinese language classroom teaching methodology to fit
the CHL context in Australia.
In addition, this study will appeal to the growing interest in Chinese Australians’
stories. The stories themselves are fascinating and absorbing because Chinese Australians
305
were never a uniform ‘heathen Chinee’ bent on flooding Australia with a low-level peasant
civilisation (Price, 1975, p. ix). Rather, they are a diverse group of people from different
areas, speaking different dialects and pursuing different aims and careers. However,
recognition of local Chinese community and heritage has largely focused on ‘gold-rush’
archaeology and immigration restrictions for much of the post-war era in Australia
(Couchman, Fitzgerald, & Macgregor, 2004). In this new millennium, there is a revitalisation
of interest in the life of Chinese Australians, among community-based amateurs, freelance
historians, and Chinese studies scholars from Australia and overseas. Moreover, many young
Chinese Australians are eager to rediscover the Chinese side of their family or personal
histories (Chan, 2004). This study will lead to a broader and deeper understanding of Chinese
Australians, with particular regard to their identity and commitment to CHL learning.
Lastly, the research is most timely. Upon the completion of the draft of the thesis, the
Australian Government released the Australia in the Asian Century White Paper in October
2012. The White Paper recognised the transformation of Asia into the economic powerhouse
of the world and stressed that Australia should seize the economic opportunities to work with
Asia. To this end, the White Paper urged more investment in Asian language education,
identifying Chinese as one of the priority languages. In recognising the economic, cultural,
and social needs to build a sound knowledge of Chinese language, the White Paper
encouraged Australians to become more Chinese literate. This will be pursued through school
education, university teaching and research, and industrial and community engagement. In
this respect, the White Paper will help shape Australia into a social field where Chinese
language has more legitimate value than at present. Resonating with the White Paper, the
research helps Australia and Australians to understand how Chinese language has become
valuable linguistic capital in contemporary Australia. In particular, the research reflects how
306
Chinese Australians negotiate their habitus of ‘Chineseness’ and capitalise on various
resources through learning Chinese. As a return on this investment, Chinese Australians will
have wider access to various forms of capital and will have a richer understanding of their
identity as being a Chinese in Australia. Consequently, Australia will benefit from this
powerful human resource that can serve the nation’s economic, cultural, social, and political
needs.
7.3 Limitations of the study
Given the practicality in research and the accessibility of resources available to research, this
study was constrained within a manageable scope. Consequently, the study, like all the other
studies, has limitations. These should be acknowledged.
The first limitation was associated with sampling. Snowball sampling was used in the
initial quantitative phase. It was argued in Chapter Four that snowball sampling was the
optimal sampling method for the current study, given the difficulty of identifying and
approaching the population under investigation. However, snowball sampling is not a
randomised method that samples the largest potential range of the demographic
characteristics of the population. This sampling method will probably net more participants
who are similar than those who are different. Consequently, it potentially reduces the
representativeness of the sample and ultimately may reduce the heterogeneity and
generalisability of the findings. These disadvantages existed in the current study.
Approximately 50% of the participants were from Brisbane because the snowball sampling
started in Brisbane. However, if the participants had been randomly sampled, Sydney should
have had the largest number of participants, followed by Melbourne and Brisbane. Another
problem with snowball sampling is that it makes it difficult to report the response rate.
307
Secondly, the pilot study in the quantitative phase was not able to check the validity of
the measurement due to limited sample size in the pilot stage. Although Cronbach’s alpha
suggested good internal consistency reliability of the measurement, this did not guarantee the
validity of the measurement. In this respect, the main study ran the risk of not being able to
achieve a reasonable level of validity, though validity of the measurement was confirmed in
CFA at a later stage. It is suggested that both reliability and validity of the measurement be
checked in the pilot phase when it is practically possible to get a sample size large enough
before major efforts are put into the main study.
Thirdly, the equations offered by SEM to calculate the scale score for the constructs
cannot be generalised to other samples. This is because the proportionally weighted factor
score of each indicator is sample-dependent in SEM. Unlike SEM, Rasch Modelling can
manage this problem. Therefore, Rasch Modelling is recommended in future research.
Next, although the explanatory mixed methods design added scope and depth to this
study, one pragmatic challenge was the longer time needed for data production and analysis
in both quantitative and qualitative ways. Good time and project management was essential to
overcome this challenge. Enhanced confidence in the trustworthiness and credibility of the
research findings, and the expanded depth and scope to the project with the application of the
mixed method design potentially offset the expanded time frame.
Another issue should also be acknowledged. The interviews in this study demanded
nuanced linguistic skill in English. This posed another challenge for the researcher, a native
Chinese speaker. When transcribing, the researcher found a number of issues that could have
been further probed at the interviews. In the interviews themselves, the researcher had to
concentrate on the literal meaning of the participants’ replies. Opportunities to explore the
deeper meanings behind participants’ replies were missed at times. Follow-up interviews
308
were done to explore these deeper meanings, which again extended the time length of the
project.
7.4 Future directions
This study has used Bourdieu’s signature notions of habitus, capital, and field to interpret
CHL practices of young Chinese Australian adults. The findings of this study supported the
claim that the habitus of ‘Chineseness’ and the investment of various forms of capital
produced practices associated with CHL, and in return, this habitus of ‘Chineseness’ was
reinforced or recaptured, and various forms of capital were generated through these CHL
practices. Besides the problems investigated in this study, there are several important and
absorbing issues emerging from this study to be addressed in future research.
Firstly, Chinese Australians never live in a closed system. Apart from habitus and
capital, there are many other sociological factors existing and emerging in social fields where
Chinese Australians interact as agents. These sociological factors, such as social class and
gender, together with habitus and capital, will have an impact on Chinese Australians’ CHL
practices. However, these sociological factors were not explored in this study. These
questions could be addressed in future research.
Secondly, the Chinese diaspora has a long history, dating back to the early 19th
century.
There are large populations of Chinese immigrants and their descendents living in different
social places, urban Australia, rural Australia, and other parts of the world. The participants in
the initial quantitative phase and the subsequent qualitative phase were sampled from urban
areas. Although an overwhelming proportion of Chinese Australians reside in urban areas, it
would be interesting to look at some cases in rural areas with perhaps less access to formal
CHL opportunities. In a global context, Chinese Australians only constitute a small body of
309
overseas Chinese. Are overseas Chinese elsewhere similar to, or different from, Chinese
Australians? It will be interesting to portray CHL practices of Chinese immigrants and their
descendents in a wider diasporic context.
Lastly, Bourdieu’s sociology is profound. His whole package of concepts and theories
will offer a substantial thinking tool when contemplating language dynamics in social,
cultural, and educational contexts. Apart from his signature notions of habitus, capital, and
field, his conceptualisation of ‘symbolic power’ and ‘symbolic violence’ would be
particularly enlightening when investigating the education and language socialisation of
indigenous people in a colonised or post-colonial context, such as the Aboriginal people in
Australia. These concepts would also be helpful when interpreting the ‘Mandarinisation’ of
the ethnic minority people, and Hong Kong and Macau people in contemporary and modern
China. In addition, these concepts could inform an examination of sociological reasons
behind the accent change of internal migrant workers in order to integrate into metropolitan
Chinese cities, such as Beijing, Shanghai, and Guangzhou. In brief, this thesis leads to
investigation of social problems of integration in a migration context, unity within a
diversified context, and the co-existence of homogeneity and heterogeneity in a globalised
context.
7.5 Overall conclusions of the study
This study started with review of the literature regarding HLLs’ identity in HL learning and
their commitment to HL learning. A rich body of literature has examined HLLs’ identity and
commitment from social psychological and poststructuralist perspectives. Related research
addressing CHLLs in particular was also reviewed. Social psychological studies have taken
an inside-out approach that tended to focus on CHLLs’ self-perceptions of their identity and
310
commitment. In contrast, poststructuralist investigations have taken an outside-in approach
that considered these issues as multiple, contradictory, ever changing, and socially
constructed through discourses. In contrast, Bourdieu’s three thinking tools of habitus, capital,
and field offer both an inside-out and an outside-in approach that can help make sense of
CHLLs’ language practices from a sociological perspective.
Beyond HL research conducted in North America, there is only a small body of
literature of relevant studies conducted in other parts of the world. Due to its idiosyncratic
colonised history, White Australia Policy, and multiculturalism, Australia is a complex social
setting for CHLLs. However, there is a scant sociological investigation taking account of both
CHLLs’ ethnic identity and their commitment to CHL learning in an Australian context.
Guided by Bourdieu’s theory, the point of departure of this study was to examine the triad of
relations amongst Chinese Australians’ ethnic identity qua habitus, their investment of
various resources as capital, and their CHL proficiency produced through language learning
practices within particular social fields.
The research design was guided by a particular philosophical position that combined a
positivist ontology, arguing that a real world exists independently of human beings’
knowledge, with a constructivist epistemology, arguing that understanding this real world is a
construction from human beings’ own perspectives and standpoint. Theoretically, Bourdieu’s
thinking through human beings’ agency and the social world’s structure fitted well in this
philosophical frame. Moreover, the methodological pluralism allowed this study to integrate
the initial quantitative investigation and the subsequent qualitative investigation to address
the research questions.
The findings from the initial quantitative phase justified Bourdieu’s theoretical
framework that agents make choices about their language practices according to the habitus
311
and capital that they have within particular fields. The analysis of the quantitative data
demonstrated that participants’ habitus of ‘Chineseness’ and their investment of various forms
of capital did contribute to their CHL proficiency. At the ontological level, Bourdieu’s
sociology can be understood as the underpinning mechanism of this emergent reality; it can
explain Chinese Australians’ CHL practices in the actual world; and the quantitative data
produced by the sampled Chinese Australians expressed this actual world in an empirical
form. However, the variance of ‘Chineseness’ and ‘capital’ only explained 62% of the
variance of ‘CHL proficiency’. The rest of the variance of ‘CHL proficiency’ was attributed
to other impact factors. At the epistemological level, habitus and capital in given fields were
the generative mechanisms that produced participants’ CHL proficiency. However, they could
not determine participants’ CHL proficiency because other social forces within fields also
contribute to this CHL proficiency.
The findings from the subsequent qualitative phase revealed the meanings associated
with CHL constructed from participants’ own perceptions. They reported developing their
habitus of ‘Chineseness’ and generating various forms of capital through CHL practice within
certain social fields. The meanings and reasoning behind CHL learning could only be
accessed through the subjective standpoint of the participants.
The findings from the quantitative phase and those from the qualitative phase
complemented and resonated with each other. The initial quantitative phase revealed that
participants’ habitus of ‘Chineseness’ and their capacity to invest in various forms of capital
generated their CHL practices that produced their CHL proficiency. In the subsequent
qualitative phase, the participants reported that they reinforced or recaptured their
‘Chineseness’ through CHL practice and they were rewarded with different forms of capital
as a return on their investment in CHL learning. As such, their CHL proficiency became
312
linguistic capital. Depending on the rules of different fields, this linguistic capital functions as
a medium to facilitate the exchange of value and forms of capital. In addition, some of the
findings from the qualitative phase conformed to those from the quantitative phase and
supported the claim that participants’ ‘Chineseness’ and their investment of capital
contributed to their CHL proficiency.
The quantitative coda demonstrated that ‘Chineseness’ and ‘capital’ were also related to
a wider range of practices associated with CHL learning. These practices included language
usage at home, years of formal CHL learning, frequency of visits to China, and self-labelling.
Furthermore, the findings from the quantitative coda also indicated the generational impacts
on the habitus of ‘Chineseness’. On the one hand, ‘Chineseness’ was durable and
transposable across space when the first-generation Chinese Australians migrated to Australia
and raised their descendents in Australia. On the other hand, ‘Chineseness’ was not
immutable across time because dispositions of this ‘Chineseness’ gradually faded away
generation after generation, accompanied by the gradual loss of CHL proficiency.
All the above findings enabled this study to talk back to the existing literature. On the
one hand, the findings of this study aligned with the existing literature that HLLs’ ethnic
identity and commitment contribute to their HL proficiency, and their HL can reinforce their
ethnic identity and their commitment to HL learning. On the other hand, the findings of this
study challenged the existing literature. The social psychological notion of motivation would
struggle to answer the questions: Why did highly motivated Adam hardly make any effort in
CHL learning when he was brought up in Indonesia? Why did En-ning, who was so
unmotivated to learn CHL when she was small, become very motivated afterwards and
decide to pursue CHL learning for her whole life? In addition, the social psychological
understanding of ethnic identity was not able to answer the question: Why did these
313
participants label themselves so differently while they were born into the same ethnic identity?
Likewise, a poststructuralist notion of investment would have trouble to explain why Bob did
not see any return on the investment of CHL learning but still committed to CHL learning
from time to time. The poststructuralist understanding of ethnic identity without a habitual
foundation is problematic: If ethnic identity is purely socially constructed, why did Adam,
growing up in an anti-Chinese social space, and En-ning, brought up in a very white area, still
keep their ‘Chineseness’, at least to a certain extent?
In contrast, Bourdieu’s triad of habitus, capital, and field offers a theoretical tool to
work with these social problems and extend social psychological and poststructuralist
understandings of HLLs’ commitment and ethnic identity. Learning Chinese did not accrue
any value in the social field in Indonesia and was not legitimised as capital, without which
Adam could hardly make any investment in CHL learning because investment demands
capital. Consequently, his CHL proficiency was low. When he moved to Australia, a place
where he considered CHL learning rewarding, he invested as much capital as possible in
CHL learning, to “catch up” in his own words. This also explained the fact that unmotivated
En-ning later became motivated to learn CHL. The social field where she was living at the
moment valued and recognised her CHL proficiency and consequently she had access to a
wider range of resources, economic, cultural, social, and symbolic, as a return on her
investment. The ethnic identity construction of Chinese Australians cannot be fully elaborated
by either the social psychological inside-out approach or the poststructuralist outside-in
approach alone. It was the habitus underpinning Adam and En-ning’s agency that made their
‘Chineseness’ durable and transposable but not immutable dispositions. It was also this
‘Chineseness’ that drove Bob to pursue CHL, consciously or unconsciously, while he did not
see any practical point of investing in CHL learning.
314
In conclusion, this study only touches a corner of sociology. Along this journey, the
researcher has been transformed from a scientist with a pure positivist ontology and
post-positivist epistemology into a scholar enlightened by the complexity and challenge of
philosophical and methodological pluralism. This different way of thinking will equip the
researcher to conduct future studies within the sociological realm. The thesis can be
concluded with a famous Chinese idiom “抛砖引玉”: Use the little to get to the big.
315
Appendices
Appendix 1: Information sheet for participation in the online survey
PARTICIPATE IN RESEARCH
Information for Prospective Participants
The following research activity has been reviewed via QUT arrangements for the conduct of research
involving human participation. If you choose to participate, you will be provided with more detailed
participant information, including who you can contact if you have any concerns.
Heritage Language for Young Chinese Australian Adults
Research Team Contacts
Guanglun Mu — PhD student Karen Dooley — supervisor
Faculty of Education Faculty of Education
Phone 0402463650 Phone 31393430
Email guanglun.mu@qut.edu.au Email k.dooley@qut.edu.au
Please contact the researcher team members to have any questions answered or if you require further
information about the project.
What is the purpose of the research?
The purpose of this project is to understand the interrelationship between Chinese Australians’ ‘Chineseness’,
their engagement in various Chinese cultural activities, and their Chinese Heritage Language.
Are you looking for people like me?
This study is looking for young Chinese Australian adults from cities, ranging in age from 18 to 35 years. You
may or may not be able to speak Chinese language. If you were born outside Australia, you will have to have
moved to Australia before the age of 13. You will be the potential participants if the above description fits
you.
What will you ask me to do?
Your participation will involve completing questions in an online survey, which may take you 15 minutes.
Are there any risks for me in taking part?
The theme of the study is related to your Chinese language learning experiences. There is no intent to elicit
sensitive or personal data. Your participation is the research is voluntary. You will be highly respected and
will be provided with detailed information about the study. You can withdraw from the study if you feel
uncomfortable during the survey.
Are there any benefits for me in taking part?
Though this project may not benefit you directly, you will help the researcher document and analyse the
language resources of young Chinese Australian adults.
I am interested – what should I do next?
If you would like to participate, please use the link below to start the online survey.
https://survey.qut.edu.au/survey/172445/5f10/ The researcher would be very grateful if you could complete the online survey at the end of September 2011.
Thank You! QUT Approval Number: 1100000165
316
Appendix 2: Information sheet for participation in the interview
PARTICIPANT INFORMATION FOR QUT RESEARCH
PROJECT
Heritage Language for Young Chinese Australian Adults
Research Team Contacts
Guanglun (Michael) Mu – PhD Candidate Dr. Karen Dooley – Supervisor
Faculty of Education Faculty of Education
Phone 0402463650 Phone 07 31383430
Email (guanglun.mu@qut.edu.au) Email (k.dooley@qut.edu.au)
Description
This interview is being undertaken as part of PhD for Guanglun (Michael) Mu. The purpose of this interview
is to understand what Chinese Heritage Language means to young Chinese Australian adults. You will be
asked questions about your Chinese language usage and learning, as well as your attitudes towards Chinese
language. The interview will take about 30 minutes at a time and place convenient for you.
Participation
Your participation in this project is voluntary. You may stop the interview at anytime if you feel
uncomfortable with the questions being asked. Your decision not to participate will in no way impact upon
your current or future relationship with QUT.
Expected benefits
It is expected that this project will not directly benefit you. You will help the researcher document and
analyse the language resources of young Chinese Australian adults.
Risks
There are no risks beyond reflecting on your Chinese learning experiences associated with your participation
in this project.
QUT provides limited free counselling for research participants of QUT projects, who may experience
discomfort or distress as result of their participation in any research. Should you wish to access to this service
please contact the Clinic Receptionist of the QUT Psychology Clinic on 3138 0999.
Confidentiality
All comments and responses will be treated confidentially. All paper-based documents will be kept securely
in a locked filing cabinet in the researcher's office. E-files and recordings will be stored on a QUT
password-protected network drive. USB drives will only be used for data transfer. Only the researcher and
the supervisory team have access to the raw data.
You have indicated that you are willing to participate in a 30-min interview. You have left your name, phone
number or email address for the researcher to contact you to arrange a convenient time and place. This
information will not be used for any other purposes. The researcher will know your identity. However, an
identity code will be used on the transcripts and pseudonyms will be used in any reports. Your identity will
only be known and available to the researcher. They will not be disclosed. Any identification will be removed
from the transcripts. The code and pseudonyms will be stored separately from your raw data. Your comments
and responses will be recorded so that they can be transcribed. You will have a chance to check and amend
the transcript as you see fit. Only the researcher and the supervisory team have access to the transcripts and
audio recordings. They will be destroyed on completion of the project.
Consent to Participate
Your signature on the consent form indicates your agreement to participate.
Questions / further information about the project
Please contact the research team members named above to have any questions answered or if you require
further information about the project.
Concerns / complaints regarding the conduct of the project
QUT is committed to researcher integrity and the ethical conduct of research projects. However, if you do
have any concerns or complaints about the ethical conduct of the project you may contact the QUT Research
Ethics Unit on +61 7 3138 5123 or email ethicscontact@qut.edu.au. The Research Ethics Unit is not
connected with the research project and can facilitate a resolution to your concern in an impartial manner.
Thank you for helping with this research project. Please keep this sheet for your information.
317
Appendix 3: The consent form
CONSENT FORM FOR QUT RESEARCH PROJECT
Heritage Language for Young Chinese Australian Adults
Research Team Contacts
Guanglun (Michael) Mu – PhD student Dr. Karen Dooley – Supervisor
Faculty of Education Faculty of Education
Phone 0402463650 Phone 07 31383430
Email (guanglun.mu@qut.edu.au) Email (k.dooley@qut.edu.au)
STATEMENT OF CONSENT
By signing below, you are indicating that you:
have read and understood the information document regarding this project
have had any questions answered to your satisfaction
understand that if you have any additional questions you can contact the research team
understand that you are free to withdraw at any time, without comment or penalty
understand that you can contact the Research Ethics Unit on +61 7 3138 5123 or email
ethicscontact@qut.edu.au if you have concerns about the ethical conduct of the project
understand that the project will include audio recording and consent to this
agree to participate in the project
Name
Signature
Date / /
Please return this sheet to the investigator.
318
Appendix 4: Interview schedule
FIRST INTERVIEW WITH INTERVIEWEE 89
Time and Date of the interview: 4pm Sydney time, 1 February, 2012
Place of the interview: Telephone interview in Room B349, Kelvin Grove
Check the audio recorders: Yes
I. Briefing
I have received the signed copy of the consent form. Thanks for that. I would also like to take
this opportunity to thank you for doing the online survey and thank you in advance for your
time to do the interview.
I will use an ID code on the transcripts and a pseudonym in any reports. Would you like to
choose your own pseudonym?
I want to assure you that there are no right or wrong answers. I have no particular agenda. I
am here to listen to and learn about your experiences.
Do you have any questions before we start the interview?
Now I will turn on the recorders.
II. Interview questions
CHL learning
First of all, I am interested in how you have gone about learning Chinese. Can you tell me
about any formal Chinese language study you have done, e.g. schools, universities,
community schools, and Confucius Institutes, etc. (probe for how long on each)?
Can you tell me about any informal Chinese learning you have pursued, e.g. with family
members, friends, independent study etc. (probe for how long on each)?
Are you still trying to improve your Chinese now? Why or why not?
Can you tell me some of your experiences or stories of how you practice your Chinese
intentionally?
CHL usage
Let’s move on to the next topic. I am interested in which language you use with whom, when
and where. Can you tell me about your language choices…
When speaking to your immediate family members / extended family members
When speaking to your (different) friends
At work
When accessing to media, e.g. internet, radio, TV, movies, music, newspaper, books,
magazines, etc.
What are your reasons behind these language choices and patterns? (Insert this question into
the above probes)
Attitudes towards CHL
Let’s talk about your attitudes towards Chinese language. Do you think your Chinese
language has benefited you in terms of …?
319
Communications with family members
Social life
Job opportunities
Chinese cultural knowledge
Credits, qualifications, awards, honours, etc.
Do you think your Chinese language will benefit you in the future? How and why?
How do you feel about (not) being able to speak Chinese?
Do you have any experiences that your Chinese language makes you feel special or different
to the others?
What do people think about your ability to use Chinese?
Family members
Friends
Colleagues
Other
If you were to have children, would you encourage them to learn another language other than
English? What is it? Why (or why not) you would encourage your children to learn Chinese?
Particular attributes
From your survey response, you are an interesting case because you have lived in several
different places. Can you tell me about your language choice when you were in Indonesia and
Singapore?
Can you tell me some of your experiences of learning Chinese when you were in Indonesia
and Singapore?
From your survey response, I know you have visited China for a few times. For what
purposes did you visit China? Can you share some of your cultural experiences in China?
How was your Chinese language helpful for your China trip?
III. Closing
Do you want to make any more comments about your Chinese language usage and learning,
as well as your attitudes towards Chinese language?
IV. Ending
Thank you again for your time. I will send you the transcript for your okay, and you are
welcome to change any wording or details to better reflect your opinions. It may take a
couple of weeks for me to get it to you.
SECOND INTERVIEW WIH INTERVIEWEE 89
Time and Date of the interview: 11am Sydney time, 26 March, 2012
Place of the interview: Telephone interview in Room B349, Kelvin Grove
Check the audio recorders: Yes
I saw your facebook photos about a speech contest. Was it related to Chinese language at all?
320
I want to learn a bit more about your life in Singapore.
How old were you when you were there?
How long did you live there?
What did you do there?
You have said in the previous interview that you tend to “force” yourself to learn Chinese and
expose yourself to Chinese language “as much as possible”. How would you interpret the
driving force behind these efforts?
I am very interested in your experience of that study tour to Shanghai in 2010.
Why did you want to join in this trip?
Is it an expensive trip? Is it worth it?
Was there anyone encouraging you to do this? Where did you get this information?
You tried to learn Mandarin in Year 12 but the school did not offer it. Can you recall your
feelings at that time?
As far as I know, not many students in Australia are interested in learning Chinese.
Where did your ideas come from?
Was there anyone encouraging you to do this?
From the previous interview, I know you have had the experience that people supposed you
can speak Chinese or even supposed you are a native Chinese speaker.
What are your reflections or feelings towards this assumption?
Did this experience affect your interest in learning Chinese in any way?
When you visited China, what was your feeling when your Chinese was understood by native
Chinese speakers?
I also learned from the last interview that you had the challenge of not being able to study
Chinese when you grew up in Jakarta. However, you do want to catch it up. I would like to
know more about this part of your experience.
Did you parents ever say anything to you at that time regarding learning Chinese?
What exactly do you want to catch up?
Why you need to catch it up?
Some of my interview participants mentioned that it is a pride to be a Chinese and speak
Chinese, and it is a shame not to be able to speak Chinese. What do you think about it?
THIRD INTERVIEW WIH INTERVIEWEE 89
Time and Date of the interview: 1:30pm Sydney time, 4 April, 2012
Place of the interview: Telephone interview in Room B349, Kelvin Grove
Check the audio recorders: Yes
When I read through the transcript of our second interview, I do have a quick question in
mind. At the end of the second interview, you indicated that being a Chinese and speaking
Chinese are more like “fit-in” for you. Can you please explain more? What exactly do you
mean by “fit-in”?
321
Bibliography
ACARA. (2011). The shape of the Australian curriculum: Languages. Sydney: Australiam
Curriculum, Assessment and Reporting Authority.
Ang, I. (2001). On not speaking Chinese: Living between Asia and the West. London:
Routledge.
Ang, I., & Stratton, J. (2001). Multilingualism in crisis: The new politics of race and national
identity in Australia. In I. Ang (Ed.), On not speaking Chinese: Living between Asia
and the West (pp. 95-111). London: Routledge.
Arnold, M. (1999). Literature & dogma: An essay towards a better apprehension of the Bible.
Cambridge: Chadwyck-Healey.
Au, T. K., Knightly, L. M., Jun, S. A., & Oh, J. S. (2002). Overhearing a language during
childhood. Psychological Science, 13(3), 238-243.
Australian Government. (2012). Australia in the Asian century white paper. Canberra:
Commonwealth of Australia.
Baker, C. (2003). Education as a site of language contact. Annual Review of Applied
Linguistics, 23, 95-112.
Baker, T. L. (1994). Doing social research (2nd ed.). New York: McGraw-Hill Inc.
Bankston, C. L., III, & Zhou, M. (1995). Effects of minority-language literacy on the
academic achievement of Vietnamese youths in New Orleans. Sociology of Education,
68(1), 1-17.
Barker, C. (2005). Cultural studies: Theory and practice. London: Sage.
Barnard, F. M. (1969). J. G. Herder on social and political culture. Cambridge: Cambridge
University Press.
Bartlett, L., & Holland, D. (2002). Theorising the space of literacy practices. Ways of
Knowing, 2(1), 10-22.
Bentler, P. M. (1990). Comparative fit indexes in structural models. Psychological Bulletin,
107(2), 238-246.
Bentler, P. M., & Chou, C.-P. (1987). Practical issues in structural modeling. sociological
Methods & Research, 16(1), 78-117.
Bernstein, B. (1973). On the classification and framing of educational knowledge. In R.
Brown (Ed.), Knowledge, education, and cultural change: Papers in the sociology of
education (pp. 363-392). London: Tavistock.
Bhatti, G. (2002). Asian children at home and at school: An ethnographic study. London:
Routledge.
Biggs, J. B. (1998). Learning from the Confucian heritage: So size doesn’t matter?
International Journal of Educational Research, 29(8), 723-738. doi:
http://dx.doi.org/10.1016/S0883-0355(98)00060-3
Biggs, J. B., & Watkins, D. A. (1996). The Chinese learner in retrospect. In D. A. Watkins & J.
B. Biggs (Eds.), The Chinese learner: Cultural, psychological, and contextual
influences (pp. 269-285). Hong Kong & Melbourne: CERC & ACER.
Binder, A. (1984). Restrictions on statistics imposed by method of measurement: Some reality,
some myth. Journal of Criminal Justice, 12(5), 467-481.
Block, D. (2002). Destabilized identities and cosmopolitanism across language and cultural
borders: Two case studies. Hong Kong Journal of Applied Linguistics, 7(2), 1-19.
Block, D. (2007). Second language identities. London; New York: Continuum.
Bourdieu, P. (1973). Cultural reproduction and social reproduction. In R. Brown (Ed.),
Knowledge, education, and cultural change: Papers in the sociology of education (pp.
71-112). London: Tavistock.
322
Bourdieu, P. (1977a). The economics of linguistic exchanges. Social Science Information,
16(6), 645-668.
Bourdieu, P. (1977b). Outline of a theory of practice. Cambridge: Cambridge University
Press.
Bourdieu, P. (1981). Structures, strategies, and the habitus. In C. C. Lemert (Ed.), French
sociology: Rupture and renewal since 1968 (pp. 86-96). New York: Columbia
University Press.
Bourdieu, P. (1985a). The genesis of the concepts of Habitus and of Field. Sociocriticism,
2(2), 11-24.
Bourdieu, P. (1985b). The social space and the genesis of groups. Theory and Society, 14(6),
723-744. doi: 10.1007/bf00174048
Bourdieu, P. (1986). Forms of capital. In J. G. Richardson (Ed.), Handbook of theory and
research for the sociology of education (pp. 241-258). New York: Greenwood Press.
Bourdieu, P. (1987). In other words: Essays towards a reflexive sociology. Stanford, CA:
Stanford University Press.
Bourdieu, P. (1988). Homo academicus. Stanford, CA: Stanford University Press.
Bourdieu, P. (1989a). Distinction: A social critique of the judgment of taste. London:
Routledge & Kegan Paul Ltd.
Bourdieu, P. (1989b). Social space and symbolic power. Sociological Theory, 7(1), 14-25.
Bourdieu, P. (1990). The logic of practice. Cambridge: Polity Press.
Bourdieu, P. (1991). Language and symbolic power. Cambridge: Polity Press.
Bourdieu, P. (1993a). The field of cultural production. New York: Columbia University Press.
Bourdieu, P. (1993b). Sociology in question. London: Sage.
Bourdieu, P. (1994). In other words: Essays towards a reflexive sociology. Cambridge: Polity
Press.
Bourdieu, P. (1996). The state nobility: Elite schools in the field of power. Cambridge: Polity
Press.
Bourdieu, P. (1998). Practical reason: On the theory of action. Cambridge: Polity Press.
Bourdieu, P. (1999a). The order of things. In P. Bourdieu (Ed.), The weight of the world:
Social suffering in contemporary society (pp. 60-76). Cambridge: Polity Press.
Bourdieu, P. (1999b). Site effects. In P. Bourdieu (Ed.), The weight of the world: Social
suffering in contemporary society (pp. 123-129). Cambridge: Polity Press.
Bourdieu, P. (1999c). Understanding. In P. Bourdieu (Ed.), The weight of the world: Social
suffering in contemporary society (pp. 607-626). Cambridge: Polity Press.
Bourdieu, P. (2000). Pascalian meditations. Cambridge: Polity Press.
Bourdieu, P. (2005). Habitus. In J. Hillier & E. Rooksby (Eds.), Habitus: A sense of place (2
ed., pp. 43-49). Aldershot, Hants, England: Ashgate Publishing Limited.
Bourdieu, P. (2011). On television. Cambridge: Polity Press.
Bourdieu, P., & Passeron, J.-C. (1990). Reproduction in education, society and culture (2nd
ed.). London: Sage.
Bourdieu, P., & Wacquant, L. J. D. (1992). An invitation to reflexive sociology. Cambridge:
Polity Press.
Brecht, R. D., & Ingold, C. W. (2002). Tapping a national resource: Heritage Languages in
the United States. ERIC Digest (ED464515). Washington, DC: ERCIC Clearinghouse
on Languages and Linguistics.
Bucholtz, M., & Hall, K. (2009). Locating identity in language. In C. Llamas & D. Watt
(Eds.), Language and identity (pp. 9-17). Ddinburgh, UK: Edinburgh University
Press.
Burrell, G., & Morgan, G. (1979). Sociological paradigms and organizational analysis.
London: Heinemann.
323
Byrne, B. M. (2001). Structural Equation Modeling with AMOS: Basic concepts, applications,
and programming. London: Lawrence Erlbaum Associates, Inc.
Campbell, R. N., & Rosenthal, J. W. (2000). Heritage Languages. In J. W. Rosenthal (Ed.),
Handbook of undergraduate Second Language education (pp. 180-201). Mahwah, NJ:
Lawrence Erlbaum Associates, Inc.
Canagarajah, A. S. (2006). TESOL at forty: What are the issues? TESOL Quarterly, 40(1),
9-34.
Carreira, M. (2004). Seeking explanatory adequacy: A dual approach to understanding the
term "Heritage Language Learner". Heritage Language Journal, 2(1), 1-25.
Case, K., & Fair, R. (1996). Principles of economics (4th ed.). Upper Saddle River, NJ:
Prentice Hall International.
Castles, S., & Davidson, A. (2000). Citizenship and migration: Globalization and the politics
of belonging. London: Macmillan.
Cattell, R. B. (1966). The scree test for the numbers of factors. Multivariate Behavioral
Research, 1, 245-276.
Cavana, R. Y., Delahaye, B. L., & Sekaran, U. (2001). Applied business research: Qualitative
and quantitative methods. Milton, QLD: John Wiley & Sons.
Celik, S. (2007). You reap what you sow: Parental investment in language learning and its
reflection in a seven-year-old's world. Taiwan Journal of TESOL, 4(1), 99-114.
Chan, H. (2004). Becoming Australian: An afterword. In S. Couchman, J. Fitzgerald & P.
Macgregor (Eds.), After the rush: Regulation, participation, and Chinese
Communities in Australia 1860-1940 (pp. 237-244). Vic., Australia: Arena Printing
and Publishing Pty Ltd.
Chao, D.-L. (1997). Chinese for Chinese Americans: A case study. Journal of the Chinese
Language Teachers Association, 32(2), 1-13.
Chen, B. D. (1966, June 1). Sweep away all the evils, People’s Daily, p. 1.
Chinen, K., & Tucker, G. R. (2005). Heritage Language development: Understanding the
roles of ethnic identity and Saturday school participation. Heritage Language Journal,
3(1), 27-59.
Cho, G. (2000). The role of Heritage Language in social interactions and relationships:
Reflections from a language minority group. Bilingual Research Journal, 24(4),
369-384.
Cho, G., Cho, K.-S., & Tse, L. (1997). Why ethnic minorities want to develop their Heritage
Language: The case of Korean-Americans. Language, Culture and Curriculum, 10(2),
106-112.
Choi, C. Y. (1975). Chinese migration and settlement in Australia. Sydney: Sydney
University Press.
Clayre, A. (1984). The heart of the dragon. Glasgow, UK: William Collins Sons & Co. Ltd.
COAG. (1994). Asian languages and Australia's economic future - A report for the Council of
Australian Governments on a proposed national Asian languages/studies strategy for
Australian schools. Brisbane: Queensland Government Printer.
Cockerham, W. C. (2005). Health lifestyle theory and the convergence of agency and
structure. Journal of Health and Social Behavior, 46(1), 51-67.
Cockerham, W. C., & Hinote, B. P. (2009). Quantifying habitus: Future directions. In K.
Robson & C. Sanders (Eds.), Quantifying theory: Pierre Bourdieu (pp. 201-210).
Breinigsville, PA, USA: Springer Science + Business Media B. V.
Cohen, J. (1988). Statistical power analysis for the behavioural sciences (2nd ed.). New York:
Academic.
Cohen, J. (1992). A power primer. Psychological Bulletin, 112(1), 155-159.
Coltman, T., Devinney, T. M., Midgley, D. F., & Venaik, S. (2008). Formative versus
324
reflective measurement models: Two applications of formative measurement. Journal
of Business Research, 61(1), 1250-1262.
Comanaru, R., & Noels, K. A. (2009). Self-determination, motivation, and the learning of
Chinese as a Heritage Language. The Canadian Modern Language Review, 66(1),
131-158.
Commonwealth Advisory Committee. (1970). The teaching of Asian languages and cultures
in Australia. Canberra: Commonwealth of Australia.
Connolly, P. (2011). Using survey data to explore preschool children's ethnic awareness and
attitudes. Journal of Early Childhood Research, 9(2), 175-187.
Costa, P. I. D. (2010). From refugee to transformer: A Bourdieusian take on a Hmong
learner's trajectory. TESOL Quarterly, 44(3), 517-541.
Costello, C. Y. (2005). Professional identity crisis: Race, class, gender, and success at
professional schools. Nashville, TN: Vanderbilt University Press.
Couchman, S., Fitzgerald, J., & Macgregor, P. (2004). After the rush: Regulation,
participation and Chinese communities in Australia 1860-1940. Vic., Australia: Arena
Printing and Publishing Pty Ltd.
Cowan, M. (1963). Humanist without portfolio: An anthology of the writings of Wilhelm von
Humboldt. Detroit: Wayne State University Press.
Coyle, D. (2002). Sex, drugs & economics: An unconventional introduction to economics.
New York: Texere.
Creswell, J. W. (2008). Educational research: Planning, conducting, and evaluating
quantitative and qualitative research (3rd ed.). Upper Saddle River, N.J:
Pearson/Merrill Prentice Hall.
Creswell, J. W. (2009). Research design: Qualitative, quantitative, and mixed methods
approaches. U.S.A.: SAGE Publications, Inc.
Creswell, J. W., & Clark, V. L. P. (2011). Designing and conducting mixed methods research.
U.S.A.: SAGE Publications, Inc.
Cummins, J. (1998). The teaching of international languages. In J. Edwards (Ed.), Language
in Canada (pp. 293-304). Cambridge: Cambridge University Press.
Cushman, J. W. (1984). A 'colonial casualty': The Chinese community in Australian
historiography. Asian Studies Association of Australia Review, 17(3), 100-113.
Danermark, B., Ekström, M., Jakobsen, L., & Karlsson, J. C. (2002). Explaining society:
Critical realism in the social sciences. London: Routledge.
De Graaf, N. D., & De Graaf, P. M. (2000). Parental cultural capital and educational
attainment in the Netherlands: A refinement of the cultural capital perspective.
Sociology of Education, 73(2), 92-111.
DEET. (1991). Australia's language: The Australian language and literacy policy.
Canberra: Australian Government Publishing Service.
Deng, Z. (2011). Confucianism, modernization and Chinese pedagogy: An introduction.
Journal of Curriculum Studies, 43(5), 561-568.
Department of Education. (1976). Report of the committee on the teaching of migrant
languages in schools. Canberra: Australian Government Publishing Service.
Diamond, J. B., Randolph, A., & Spillane, J. P. (2004). Teachers' expectations and sense of
responsibility for student learning: The importance of race, class, and organizational
habitus. Anthropology & Education Quarterly, 35(1), 75-98. doi:
10.1525/aeq.2004.35.1.75
DiMaggio, P. (1982). Cultural capital and school success: The impact of status culture
participation on the grades of U.S. high school students. American Sociological
Review, 47(2), 189-201.
Dörnyei, Z., & Skehan, P. (2003). Individual differences in second language learning. In C.
325
Doughty & M. H. Long (Eds.), Handbook of second language acquisition (pp.
589-630). Malden, MA: Blackwell.
Dovey, K. (2010). Becoming places: Urbanism, architecture, identity and power. London:
Routledge.
Draper, J., & Hicks, J. (2000). Where we've been; what we've learned. In J. B. Webb & B.
Miller (Eds.), Teaching Heritage Language Learners: Voices from the classroom (pp.
15-35). Yonkers, NY: ACTFL.
Duhachek, A., Coughlan, A. T., & Iacobucci, D. (2005). Results on the standard error of the
Coefficient Alpha Index of reliability. Marketing Science, 24(2), 294-301.
Dumais, S. A. (2002). Cultural capital, gender, and school success: The role of habitus.
Sociology of Education, 75(1), 44-68.
Durbin, J., & Watson, G. S. (1950). Testing for Serial Correlation in Least Squares
Regression, I. Biometrika, 37(3/4), 409-428.
Durbin, J., & Watson, G. S. (1951). Testing for Serial Correlation in Least Squares
Regression, II. Biometrika, 38(1/2), 159-179.
Durkheim, E. (1938). L'évolution pedagogique en France. Paris: Alcan.
Durkheim, E. (1964). The rules of sociological method. New York: Free Press.
Ellis, R. (1997). Second Language Acquisition. Oxford: Oxford University Press.
Erikson, E. (1968). Identity: Youth and crisis. New York: Norton.
Feuerverger, G. (1989). Jewish-Canadian ethnic identity and non-native language learning: A
social psychological study. Journal of Multilingual and Multicultural Development,
10(4), 327-357.
Feuerverger, G. (1991). University students' perceptions of Heritage Language learning and
ethnic identity maintenance. The Canadian Modern Language Review, 47(4),
660-677.
Field, A. (2009). Discovering statistics using SPSS (3rd ed.). London: Sage.
Fishman, J. A. (1964). Language maintenance and language shift as a field of inquiry.
Linguistics, 9, 32-70.
Fishman, J. A. (1977). Language and ethnicity. In H. Giles (Ed.), Language, ethnicity and
intergroup relations (pp. 15-57). London: Academic Press.
Fishman, J. A. (1989). The historical and social contexts of an inquiry into language and
maintenance efforts. In J. A. Fishman (Ed.), Language and ethnicity in minority
sociolinguistic perspective (pp. 21-33). Clevedon: Multilingual Matters.
Fishman, J. A. (1991). Reversing language shift. Clevedon: Multilingual Matters.
Fishman, J. A. (2001). 300-plus years of Heritage Language education in the United States. In
J. K. Peyton, D. A. Ranard & S. McGinnis (Eds.), Heritage Languages in America:
Preserving a national resource (pp. 29-36). McHenry, IL: Centre for Applied
Linguistics / Delta Systems.
Fornell, C., & Larcker, D. F. (1981). Evaluating Structural Equation Models with
unobservable variables and measurement error. Journal of Marketing Research, 18,
39-50.
Fraenkel, J. R., & Wallen, N. E. (1996). How to design and evaluate research in education
(3rd ed.). New York: McGraw-Hill.
Francis, B., Archer, L., & Mau, A. (2009). Language as capital, or language as identity?
Chinese complementary school pupils' perspectives on the purposes and benefits of
complementary schools. British Educational Research Journal, 35(4), 519-538.
Galbally, F. (1978). Migrant services and programs: Report of the review of post-arrival
programs and services for migrants (Galbally Report). Canberra: Australian
Government Publishing Services.
Gardner, R. C. (1968). Attitudes and motivation: Their role in Second Language Acquisition.
326
TESOL Quarterly, 2(3), 141-150.
Gardner, R. C. (1985). Social psychology and language learning: The role of attitudes and
motivation. London: Edward Arnold.
Gardner, R. C., & Lambert, W. E. (1959). Motivational variables in Second Language
Acquisition. Canadian Journal of Psychology, 13(4), 266-272.
Gardner, R. C., & Lambert, W. E. (1972). Attitudes and motivation in Second Language
Learning. Rowley, MA: Newbury House.
Garnham, N., & Williams, R. (1993). Pierre Bourdieu and sociology of culture: An
introduction. In R. Collins, J. Curran, N. Garnham, P. Scannell, P. Schlesinger & C.
Sparks (Eds.), Media, culture & society: A critical reader (pp. 116-130). London:
Sage.
Gibbs, J. T., & Hines, A. M. (1992). Negotiating ethnic identity: Issues for Black-White
biracial adolescents. In M. P. Root (Ed.), Racially mixed people in America (pp.
223-238). Newbury Park, CA: Sage.
Giles, H., Bourhis, R., & Taylor, D. (1977). Toward a theory of language in ethnic group
relations. In H. Giles (Ed.), Language, ethnicity and intergroup relations (pp. 32-65).
London: Academic Press.
Giles, H., & Johnson, P. (1987). Ethnolinguistic identity theory. International Journal of the
Sociology of Language, 68, 69-99.
Gogolin, I. (2002). Linguistic and cultural diversity in Europe: A challenge for educational
research and practice. Eruopean Educational Research Journal, 1(1), 123-138.
Graham, J. M. (2006). Congeneric and (essentially) Tau-equivalent estimates of score
reliability. Educational and Psychological Measurement, 66(6), 930-944.
Green, S. B. (1991). How many subjects does it take to do a regression analysis? Multivariate
Behavioral Research, 26(3), 499-510.
Groves, R. M., Fowler, F. J., Couper, M. P., Lepkowski, J. M., Singer, E., & Tourangeau, R.
(2004). Survey methodology. New Jersey: John Wiley & Sons, Inc., Hoboken.
Guba, E. G., & Lincoln, Y. S. (1994). Competing paradigms in qualitative research. In N. K.
Denzin & Y. S. Lincoln (Eds.), Handbook of qualitative research (pp. 105-117).
Thousand Oaks, CA: Sage.
Hair, J. F., Black, B. J., Babin, B. J., Anderson, R. E., & Tatham, R. L. (2006). Multivariate
data analysis (6th ed.). Upper Saddle River, NJ: Pearson Prentice Hall.
Hall, C. I. (1992). Please choose one. In M. P. Root (Ed.), Racially mixed people in America
(pp. 250-264). Newbury Park, CA: Sage.
Hancock, A. (2006). Attitudes and approaches to literacy in Scottish Chinese families.
Language and Education, 20(5), 355 - 373.
Hancock, G. R., & Mueller, R. O. (2001). Rethinking construct reliability within latent
variables systems. In R. Cudeck, S. D. Toit & D. Sörbom (Eds.), Structural Equation
Modeling: Present and future- A festschrift in honor of Karl Jöreskog (pp. 195-216).
Lincolnwood: Scientific Software International.
Haralambos, M., & Holborn, M. (2004). Sociology: Themes and perspectives (6th ed.).
London: Harper Collins Publishers Limited.
Hase, P. (1995). The alliance of ten: Settlement and politics in the Sha Tau Kok Area. In D.
Faure & H. F. Siu (Eds.), Down to earth: The territorial bond in South China (pp.
123-160). Standford, California: Standford University Press.
He, A. W. (2004). Identity construction in Chinese Heritage Language classes. Pragmatics,
14(2/3), 199-216.
He, A. W. (2005). Discipline, directives, and deletions: Grammar and interaction in Chinese
heritage language classes. In C. Holten & J. Frodesen (Eds.), The power of context in
language teaching and learning: A Festschrift for Marianne Celce-Murcia (pp.
327
115-126). Boston: Thomson Heinle.
He, A. W. (2006). Toward an identity theory of the development of Chinese as a Heritage
Language. Heritage Language Journal, 4(1), 1-28.
He, A. W. (2008). Chinese as a Heritage Language: An introduction. In A. W. He & Y. Xiao
(Eds.), Chinese as a Heritage Language: Fostering rooted world citizenry (pp. 1-12).
Hololulu: University of Hawai’i Press.
Hinton, L. (1999). Involuntary language loss among immigrants: Asian-American linguistic
autobiographies. (Report No. EDO-FL-99-10). Retrieved from CAL Center for
Applied Linguistics website: http://www.cal.org/resources/digest/involuntary.html.
Ho, D. Y. F., Xie, W., Liang, X., & Zeng, L. (2012). Filial piety and traditional Chinese
values: A study of high and mass cultures. PsyCh Journal, 1(1), 40-55. doi:
10.1002/pchj.6
Ho, R. (2006). Handbook of univariate and multivariate data analysis and interpretation with
SPSS. Boca Raton: Chapman & Hall/CRC.
Hodge, B., & O’Carroll, J. (2006). Borderwork in multicultural Australia. Sydney: Allen &
Unwin.
Holland, D., Lachicotte, W., Skinner, D., & Cain, D. (1998). Identity and agency in cultural
worlds. Cambridge, MA: Harvard University Press.
Holmes, D., Hughes, K., & Julian, R. (2007). Australian sociology: A changing society.
Frenchs Forest, N.S.W.: Pearson Education Australia.
Horvat, E. M., & Antonio, A. L. (1999). "Hey, those shoes are out of uniform": African
American girls in an elite high school and the importance of habitus. Anthropology &
Education Quarterly, 30(3), 317-342.
Hurtado, A., & Gurin, P. (1995). Ethnic identity and bilingualism attitudes. In A. M. Padilla
(Ed.), Hispanic psychology: Critical issues in theory and research (pp. 89-103).
Thousand Oaks, CA: Sage.
Imbens-Bailey, A. (1996). Ancestral language acquisition: Implications for aspects of ethnic
identity among Armenian American children and adolescents. Journal of Language
and Social Psychology, 15(4), 422-443.
Ingram, D. E. (2000, 18 August). Language policy and language education in Australia.
Paper presented at the Akita University, Japan.
Jackson, P. (2009). Pierre Bourdieu. In J. Edkins & N. Vaughan-Williams (Eds.), Critical
theorists and international relations (pp. 102-113). New York: Routledge.
Jenkins, R. (1992). Pierre Bourdieu. London and New York: Routledge.
Jones, P. (2004). What happened to Australia's Chinese between the World Wars? In S.
Couchman, J. Fitzgerald & P. Macgregor (Eds.), After the rush: Regulation,
participation, and Chinese communities in Australia 1860-1940 (pp. 217-236). Vic.,
Australia: Arena Printing and Publishing Pty Ltd.
Jones, P. (2005). Chinese-Australian journeys: Records on travel, migration and settlement,
1860-1975. Canberra: National Archives.
Jöreskog, K. G., & Sörbom, D. (1996). LISERAL 8 user's reference guide. Mooresville, IN:
Scientific Software International.
Kaiser, H. F. (1960). An application of electronic computers to factor analysis. Educational
and Psychological Measurement, 20(1), 141-151.
Kaiser, H. F. (1974). An index of factorial simplicity. Psychometrika, 39(1), 31-36.
Ke, C. (1998). Effects of language background on the learning of Chinese characters among
foreign languages students. Foreign Language Annals, 31(1), 91-102.
Kelley, K., & Lai, K. (2011). Accuracy in parameter estimation for the Root Mean Square
Error of Approximation: Sample size planning for narrow confidence intervals.
Multivariate Behavioral Research, 46(1), 1-32. doi: 10.1080/00273171.2011.543027
328
Kiang, L. (2008). Ethnic self-labeling in young American adults from Chinese backgrounds.
Journal of Youth and Adolescence, 37(1), 97-111.
Kim, K. H. (2005). The relation among fit indexes, power, and sample size in Structural
Equation Modeling. Structural Equation Modeling: A Multidisciplinary Journal,
12(3), 368-390. doi: 10.1207/s15328007sem1203_2
Kline, P. (1999). The handbook of psychological testing (2nd ed.). London: Routledge.
Kramsch, C. (2008). Pierre Bourdieu: A biographical memoir. In A. L. J. Albright (Ed.),
Bourdieu and literacy education (pp. 33-49). London: Routledge.
Kvale, S. (1996). Interviews: An introduction to qualitative research interviewing. London:
Sage.
Labov, W. (1972). Language in the inner city: Studies in the Black English vernacular.
Philadelphia: University of Pennsylvania Press.
Lai, K. L. (2008). An introduction to Chinese philosophy. Cambridge: Cambridge University
Press.
Landis, R. S., Beal, D. J., & Tesluk, P. E. (2000). A Comparison of approaches to forming
composite measures in Structural Equation Models. Organizational Research
Methods, 3(2), 186-207. doi: 10.1177/109442810032003
Lane, J. F. (2000). Pierre Bourdieu: A critical introduction. London: Pluto Press.
Lao, C. (2004). Parents' attitudes toward Chinese-English bilingual education and
Chinese-language use. Bilingual Research Journal, 28(1), 99-121.
Lee, J. S. (2002). The Korean language in America: The role of cultural identity in Heritage
Language learning. Language, Culture and Curriculum, 15(2), 117-133.
Lee, W. O. (1996). The cultural context for Chinese learners: Conceptions of learning in the
Confucian tradition. In D. A. Watkins & J. B. Biggs (Eds.), The Chinese Learner (pp.
45-67). Hong Kong: Comparative Education Research Centre.
Lehman, R. (1991). Statistics and research design in the behavioral sciences. Belmont,
California: Wadsworth Publishing Company.
Levesque, R. J. R. (2007). The ethnicity of adolescent research. Journal of Youth and
Adolescence, 36(4), 375-389.
Levinson, D. J. (1978). The seasons of a man's life. New York: Ballantine.
Li, D. (2005, 25-29 July). Attitudes, motivations and identities in learning Chinese as a
Heritage Language. Paper presented at the 14th World Congress of Applied
Linguistics, Madison, WI.
Li, D., & Duff, P. A. (2008). Issues in Chinese Heritage Language education and research at
the postsecondary level. In A. W. He & Y. Xiao (Eds.), Chinese as a Heritage
Language: Fostering rooted world citizenry (pp. 13-36). Honolulu: University of
Hawaii's Press, National Foreign Language Resource Center.
Li, W. (1994). Three generations, two languages, one family: Language choices and
language shift in a Chinese community in Britain. Clevedon, England: Multilingual
Matters.
Li, W., & Zhu, H. (2011). Changing hierarchies in Chinese language education for the British
Chinese learners. In L. Tsung & K. Cruickshank (Eds.), Teaching and learning
Chinese in global contexts: Multimodality and literacy in the new media age (pp.
11-27). London & New York: Continuum International Publishing Group.
Lin, H. (2004). The development of Hokkien and its relationship with the cultural identity in
Fujian Province in China. China: Southeast Research.
Lingard, B., Rawolle, S., & Taylor, S. (2005). Globalising policy sociology in education:
Working with Bourdieu. Journal of Education Policy, 20(6), 759-777.
Lo-Philip, S. W.-Y. (2010). Towards a theoretical framework of Heritage Language literacy
and identity processes. Linguistics and Education: An International Research Journal,
329
21(4), 282-297
Lo Bianco, J. (1987). National policy on languages. Canberra: Australian Government
Publishing Service.
Lo Bianco, J. (1990). Making language policy: Australia's experience. In R. B. Baldauf & A.
Luke (Eds.), Language planning and education in Australasia and the South Pacific
(pp. 47-79). Philadelphia: Multilingual Matters.
Lo Bianco, J. (2000). Making languages an object of public policy. Agenda, 7(1), 47-61.
Lo Bianco, J. (2011). Chinese: The gigantic up-and-comer. In L. Tsung & K. Cruickshank
(Eds.), Teaching and learning Chinese in global contexts: Multimodality and literacy
in the new media age (pp. xiii-xxiv). London & New York: Continuum International
Publishing Group.
Lodico, M. G., Spaulding, D. T., & Voegtle, K. H. (2006). Methods in educational research:
From theory to practice. San Francisco: Jossey-Bass.
Lokan, J., Ford, P., & Greenwood, L. (1997). Maths & science on the line: Australian middle
primary students' performance in the third international mathematics and science
study. Melbourne: Australia Council for Educational Research.
Louie, A. (2004). Chineseness across borders: Renegotiating Chinese identities in China and
the United States. Durham and London: Duke University Press.
Louie, K. (2008). Defining modern Chinese culture. In K. Louie (Ed.), The Cambridge
companion to modern Chinese culture (pp. 1-19). Cambridge: Cambridge University
Press.
Lu, X., & Li, G. (2008). Motivation and achievement in Chinese language learning: A
comparative analysis. In A. W. He & Y. Xiao (Eds.), Chinese as a Heritage Language:
Fostering rooted world citizenry (pp. 89-108). Hololulu: The University of Hawaii
Press.
Luke, A. (2009). Race and language as capital in school: A sociological template for language
education reform. In R. Kubota & A. M. Y. Lin (Eds.), Race, culture and identities in
second language education: Exploring critically engaged practice (pp. 286-308).
London: Routledge.
Luke, A. (2010). Documenting reproduction and inequality: Revisiting Jean Anyon's “social
class and school knowledge”. Curriculum Inquiry, 40(1), 167-182. doi:
10.1111/j.1467-873X.2009.00474.x
Luo, S.-H., & Wiseman, R. L. (2000). Ethnic language maintenance among Chinese
immigrant children in the United States. International Journal of Intercultural
Relations, 24(3), 307-324.
Lynch, A. (2003). The relationship between Second and Heritage Language Acquisition:
Notes on research and theory building. Heritage Language Journal, 1, 1-18.
Macknight, C. C. (1976). The voyage to Marege' : Macassan trepangers in Northern
Australia Melbourne: Melbourne University Press.
Mactavish, J. B., & Schleien, S. J. (2000). Beyond qualitative and quantitative data linking:
An example from a mixed method study of family recreation. Therapeutic Recreation
Journal, 34(2), 154-163.
Marks, G. N. (2009). The influence of cultural capital on educational and early labor market
outcomes of young people in Australia. In K. Robson & C. Sanders (Eds.),
Quantifying theory: Pierre Bourdieu (pp. 89-104). Breinigsville, PA, USA: Springer
Science + Business Media B. V.
Maton, K. (2008). Habitus. In M. Grenell (Ed.), Pierre Bourdieu: Key concepts (pp. 49-66).
Stocksfield: Acumen.
McClelland, K. (1990). Cumulative disadvantage among the highly ambitious. Sociology of
Education, 63(2), 102-121.
330
McGinnis, S. (2005). More than a silver bullet: The role of Chinese as a Heritage Language
in the United States. The Modern Language Journal, 89(4), 592-594.
McGinnis, S. (2008). From mirror to compass: The Chinese Heritage Language education
sector in the United States. In D. M. Brinton, O. Kagan & S. Bauckus (Eds.), Heritage
Language education: A new field emerging (pp. 229-242). New York: Routledge.
McLaren, A. (2011). Asian languages enrolments in Australian higher education 2008-09.
Canberra: The Asian Association of Australia.
Menard-Warwick, J. (2005). Intergenerational trajectories and sociopolitical context: Latina
immigrants in adult ESL. TESOL Quarterly, 39(2), 165-185.
Merriam, S. B. (1998). Qualitative research and case study applications in education. San
Francisco, CA: Jossey-Bass.
Mertens, D. M. (2005). Research and evaluation in education and psychology: Integrating
diversity with quantitative, qualitative, and mixed methods. U.S.A.: Sage.
Mertler, C. A., & Charles, C. M. (2005). Introduction to educational research (5th ed.).
U.S.A.: Pearson Education, Inc.
Mu, G. M. (2012, 24-27 November). Does ‘habitus’ count in Chinese Australians’
mathematics achievement? Paper presented at the 2nd International STEM in
Education Conference, Beijing.
Nardi, P. M. (2006). Doing a survey research: A guide to quantitative methods (2nd ed.).
U.S.A.: Pearson Education Inc.
Noels, K. A. (2005). Orientations to learning German: Heritage Language learning and
motivational substrates. The Canadian Modern Language Review, 62, 285-312.
Norton, B. (1994, March). Language learning, social identity, and immigrant women. Paper
presented at the Annual Meeting of the Teachers of English to Speaks of Other
Languages, Baltimore.
Norton, B. (1995). Social identity, investment, and language learning. TESOL Quarterly,
29(1), 9-31.
Norton, B. (2000). Identity and language learning: Gender, ethnicity, and educational change.
Harlow, England: Pearson Education.
Ochs, E. (1993). Constructing social identity: A language socialization perspective. Research
on Language and Social Interaction, 26, 287-306.
Oh, J. S., & Fuligni, A. J. (2010). The role of Heritage Language development in the ethnic
identity and family relationships of adolescents from immigrant backgrounds. Social
Development, 19(1), 202-220.
Ong, A. (1999). Flexible citizenship: The cultural logistics of transnationality. Durham, N.C.:
Duke University Press.
Ong, A., Dominguez, V. R., Friedman, J., Schiller, N. G., Stolcke, V., Wu, D. Y. H., & Ying,
H. (1996). Cultural citizenship as subject-making: Immigrants negotiate racial and
cultural boundaries in the United States [and comments and reply]. Current
Anthropology, 37(5), 737-762.
Orton, J. (2008). Chinese language education in Australian schools. Melbourne: Melbourne
Graduate School of Education, The University of Melbourne.
Pahl, K. (2008). Tracing habitus in texts: Narratives of loss, displacement and migration in
homes. In J. Albright & A. Luke (Eds.), Bourdieu and literacy education (pp.
187-208). London: Routledge.
Pao, D. L., Wong, S. D., & Teuben-Rowe, S. (1997). Identity formation for mixed-heritage
adults and implications for educators. TESOL Quarterly, 31(3), 622-631.
Patton, M. Q. (1990). Qualitative evaluation and research methods. Newbury Park, CA:
Sage.
Pavlenko, A. (2002). Poststructuralist approaches to the study of social factors in second
331
language learning and use. In V. Cook (Ed.), Portraits of the L2 user (pp. 275-302).
Clevedon, Buffalo, Toronto, Sydney: Multilingual Matters Ltd.
Pennycook, A. (2003). Global Englishes, rip slime, and performativity. Journal of
Sociolinguistics, 7(4), 513-533.
Peräkylä, A. (2004). Reliability & validity in research based on naturally ocurring social
interaction. In D. Silverman (Ed.), Qualitative research: Theory, method and practice
(2 ed., pp. 283-304). London: Sage.
Peyton, J. K., Ranard, D., & McGinnis, S. (2001). Charting a new course: Heritage Language
education in the United States. In J. K. Peyton, D. Ranard & S. McGinnis (Eds.),
Heritage Languages in America: Preserving a national resource (pp. 3-26).
Washington, DC: Center for Applied Linguistics.
Phinney, J. S., & Ong, A. D. (2007). Conceptualisation and measurement of ethnic identity:
Current status and future directions. Journal of Counseling Psychology, 54(3),
271-281.
Phinney, J. S., Romero, I., Nava, M., & Huang, D. (2001). The role of language, parents, and
peers in ethnic identity among adolescents in immigrant families. Journal of Youth
and Adolescence, 30(2), 135.
Price, C. A. (1975). Foreword. In C. Y. Choi (Ed.), Chinese migration and settlement in
Australia (pp. ix-x). Sydney: Sydney University Press.
Purdey, J. (2006). Anti-Chinese violence in Indonesia, 1996-1999. Honolulu: University of
Hawai'i Press.
Reay, D. (1995). 'They employ cleaners to do that': Habitus in the primary classroom. British
Journal of Sociology of Education, 16(3), 353-371.
Reay, D. (2004). 'It's all becoming a habitus': Beyond the habitual use of habitus in
educational research. British Journal of Sociology of Education, 25(4), 431-444.
Root, M. P. (1992). Within, between, and beyond race. In M. P. Root (Ed.), Racially mixed
people in America (pp. 3-11). Newbury Park, CA: Sage.
Roscigno, V. J., & Ainsworth-Darnell, J. W. (1999). Race, cultural capital, and educational
resources: Persistent inequalities and achievement returns. Sociology of Education,
72(3), 158-178.
Rosenthal, D. A., & Feldman, S. S. (1992). The nature and stability of ethnic identity in
Chinese youth: Effects of length of residence in two cultural contexts. Journal of
Cross-Cultural Psychology, 23(2), 214-227.
Rowe, K. (2002, October 11). The measurement of latent and composite variables from
multiple items or indicators: Applications in performance indicator systems. Paper
presented at the Royal Molbourne Institute of Technology Statistics Seminar Series,
Melbourne.
Rowsell, J. (2008). Improvising on artistic habitus: Sedimenting identity into art. In J.
Albright & A. Luke (Eds.), Bourdieu and literacy education (pp. 233-251). London:
Routledge.
Rummens, J. A. (2003). Conceptualizing identity and diversity: Overlaps, intersections, and
processes. Canadian Ethnic Studies Journal, 35(3), 10-25.
Ryan, F., Coughlan, M., & Cronin, P. (2009). Interviewing in qualitative research: The
one-to-one interview. International Journal of Therapy & Rehabilitation, 16(6),
309-314.
Sadovnik, A. R. (2011). Theory and research in the sociology of education. In A. R. Sadovnik
(Ed.), Sociology of education: A critical reader (pp. 3-22). New York: Routledge.
Salili, F., Chiu, C. Y., & Lai, S. (2001). The influence of culture and context on students’
achievement orientations. In F. Salili, C. Y. Chiu & Y. Y. Hong (Eds.), Student
motivation: The culture and context of learning (pp. 221-247). New York: Plenum.
332
Shen, H. H. (2003). A comparison of written Chinese achievement among Heritage Learners
in homogeneous and heterogeneous groups. Foreign Language Annals, 36(2),
258-266.
Shin, S. J. (2010). "What about me? I'm not like Chinese but I'm not like American”: Heritage
Language learning and identity of mixed-heritage adults. Journal of Language,
Identity & Education, 9(3), 203-219.
Silverman, D. (1993). Interpreting qualitative data: Methods for analyzing talk, text and
interaction. London: Sage.
Smolicz, J. J. (1981). Language as core value of culture. In H. B. Beardsmore (Ed.), Elements
of bilingual theory (pp. 104-124). Brussels: Vrije Universiteit.
Stevens, J. P. (2002). Applied multivariate statistics for the social sciences (4th ed.). Hillsdale,
NJ: Erlbaum.
Studenmund, A. H., & Cassidy, H. J. (1987). Using economietrics: A practical guide. Boston:
Little Brown.
Sturak, K., & Naughten, Z. (2010). The current state of Chinese, Japanese, Indonesian and
Korean language education in Australian schools: Four languages, four stories
(ISBN: 978 1 74200 105 0). Carlton South, Vic.: Education Services Australia
Retrieved from http://www.deewr.gov.au/Schooling/NALSSP/Pages/Resources.aspx.
Sullivan, A. (2001). Cultural capital and educational attainment. Sociology, 35(4), 893-912.
Sullivan, A. (2002). Bourdieu and education: How useful is Bourdieu's theory for researchers?
The Netherlands' Journal of Social Sciences, 38(2), 144-166.
Sun, H., Hu, Z., & Huang, X. (2007). Languages in China. Beijing: The Commercial Press.
Tabachnick, B. G., & Fidell, L. S. (2007). Using multivariate statistics (5th ed.). Boston:
Allyn and Bacon.
Tajfel, H. (1981). Human groups and social categories. Cambridge: Cambridge University
Press.
Tan, S.-H. (2008). Modernizing Confucianism and ‘new Confucianism’. In K. Louie (Ed.),
The Cambridge companion to modern Chinese culture (pp. 135-154). Cambridge:
Cambridge University Press.
Tashakkori, A., & Teddlie, C. (2003). Handbook of mixed methods in social and behavioral
research. Thousand Oaks, CA: Sage.
Thomson, P. (2008). Field. In M. Grenfell (Ed.), Pierre Bourdieu: Key concepts (pp. 67-81).
Stocksfield, UK: Acumen.
Tod, A. (2006). Interviewing. In K. Gerrish & A. Lacey (Eds.), The research process in
nursing (pp. 337-352). Oxford: Blackwell Publishing.
Tse, L. (2001). Why don’t they learn English? Separating fact from fallacy in the U.S.
language debate. New York: Teacher’s College Press.
Tsung, L., & Cruickshank, K. (2011). Teaching and learning Chinese: A research agenda. In L.
Tsung & K. Cruickshank (Eds.), Teaching and learning Chinese in global contexts:
Multimodality and literacy in the new media age (pp. 213-224). London & New York:
Continuum International Publishing Group.
Valdés, G. (2001). Heritage Language students: Profiles and personalities. In J. K. Peyton, D.
A. Ranard & S. McGinnis (Eds.), Heritage Languages in America: Preserving a
national resource (pp. 29-36). McHenry, IL: Centre for Applied Linguistics / Delta
Systems.
Van Deusen-Scholl, T. (2003). Toward a definition of Heritage Language: Sociopolitical and
pedagogical considerations. Journal of Language, Identity & Education, 2(3),
211-230.
Veenstra, G. (2009). Transmutations of capitals in Canada: A ‘social space’ approach. In K.
Robson & C. Sanders (Eds.), Quantifying theory: Pierre Bourdieu (pp. 61-74).
333
Breinigsville, PA, USA: Springer Science + Business Media B. V.
Von Glasersfeld, E. (2001). The radical constructivist view of science. Foundation of Science,
6(1-3), 31-43.
Vryonides, M. (2007). Social and cultural capital in educational research: Issues of
operationalisation and measurement. British Educational Research Journal, 33(6),
867-885.
Wacquant, L. J. D. (1989). Towards a reflexive sociology: A workshop with Pierre Bourdieu.
Sociological Theory, 7(1), 26-63.
Walonick, D. (1993). StatPac gold IV: Marketing research and survey edition. Minneapolis,
MN: StatPac, Inc.
Wang, S. C. (2007). Building societal capital: Chinese in the US. Language Policy, 6(1),
27-52. doi: 10.1007/s10993-006-9043-2
Watson, J. L. (1975). Emigration and the Chinese lineage: The Mans in Hong Kong and
London. Berkeley: University of California Press.
Webb, J., Schirato, T., & Danaher, G. (2002). Understanding Bourdieu. Australia: Allen &
Unwin.
Webb, J. B., & Miller, B. L. (2000). Teaching Heritage Language Learners: Voices from the
classroom. Yonkers, NY: ACTFL.
Weedon, C. (1997). Feminist practice and poststructuralist theory (2nd ed.). Oxford:
Blackwell.
Weerakkody, N. (2010). Roots and routes. In M. Kumar, H. D'Cruz & N. Weerakkody (Eds.),
Where are you from? Voices in transition (pp. 25-37). Altona, Vic.: Common Ground.
Weger-Guntharp, H. (2006). Voices from the margin: Developing a profile of Chinese
Heritage Language Learners in the FL classroom. Heritage Language Journal, 4(1),
29-46.
Wen, X. (1997). Motivation and language learning with students of Chinese. Foreign
Language Annals, 30(2), 235-251. doi: 10.1111/j.1944-9720.1997.tb02345.x
Wen, X. (2011). Chinese language learning motivation: A comparative study of Heritage and
non-Heritage Learners. Heritage Language Journal, 8(3), 41-66.
Williams, R. (1988). Keywords: A vocabulary of culture and society. London: Fontana Press.
Wong-Fillmore, L. (1991). When learning a second language means losing the first. Early
Childhood Research Quarterly, 6(3), 323-346.
Wong, K. F., & Xiao, Y. (2010). Diversity and difference: Identity issues of Chinese Heritage
Language Learners from dialect backgrounds. Heritage Language Journal, 7(2),
153-187.
Woodward, K. (1997). Identity and difference. London: Sage.
Woodward, K. (2003). Understanding identity. London: Arnold.
Xi, X., & Gao, C. M. (2005). A brief history of the Cultural Revolution (2nd ed.). Beijing:
Chinese Communist Party History Press.
Xiao, Y. (2006). Heritage Learners in the Chinese language classroom: Home background.
Heritage Language Journal, 4(1), 47-56.
Xiao, Y. (2008). Home literacy environment in Chinese as a Heritage Language. In A. W. He
& Y. Xiao (Eds.), Chinese as a Heritage Language: Fostering rooted world citizenry
(pp. 151-166). Hololulu: University of Hawai’i Press, National Foreign Language
Resource Center.
Xing, J. Z. (2006). Teaching and learning Chinese as a Foreign Language: A pedagogical
grammar. Hong Kong: Hong Kong University Press.
Yamamoto, Y., & Brinton, M. C. (2010). Cultural capital in East Asian educational systems:
The case of Japan. Sociology of Education, 83(1), 67-83.
Yang, J. S. R. (2003). Motivational orientations and selected learner variables of East Asian
334
language learners in the United States. Foreign Language Annals, 36(1), 44-56. doi:
10.1111/j.1944-9720.2003.tb01931.x
Yang, K. S. (1986). Chinese personality and its change. In M. H. Bond (Ed.), The psychology
of the Chinese people (pp. 106-170). Hong Kong: Oxford University Press.
Yuan, K. H., Marshall, L. L., & Bentler, P. M. (2002). A unified approach to exploratory
factor analysis with missing data, nonnormal data, and in the presence of outliers.
Psychometrika, 67(1), 95-122.
Yurdugul, H. (2008). Minimum sample size for Cronbach's coefficient alpha: A Monte-Carlo
study. H. U. Journal of Education, 35, 397-405.
Zacher, J. (2008). Social hierarchies and identity politics. In J. Albright & A. Luke (Eds.),
Bourdieu and literacy education (pp. 252-278). London: Routledge.
Zhang, D., & Slaughter-Defoe, D. T. (2009). Language attitudes and Heritage Language
maintenance among Chinese immigrant families in the USA. Language, Culture and
Curriculum, 22(2), 77-93.
Zhang, J. (2009). Mandarin maintenance among immigrant children from the People’s
Republic of China: An examination of individual networks of linguistic contact.
Language, Culture and Curriculum, 22(3), 195-213.
Zhang, Q. Y. (2009). Confucianism for today. Beijing: Peking University Press.
Zhao, D., & Singh, M. (2011). Why do Chinese-Australian students outperform their
Australian peers in mathematics: A comparative case study. International Journal of
Science and Mathematics Education, 9(1), 69-87. doi: 10.1007/s10763-010-9214-7
Zumbo, B. D., & Zimmerman, D. W. (1993). Is the selection of statistical methods governed
by level of measurement? Canadian Psychology, 34(4), 390-399.