192 5 4MB
English Pages 214 [216] Year 2011
Early Child Cantonese
Studies on Language Acquisition 42
Editor Peter Jordens
De Gruyter Mouton
Early Child Cantonese Facts and Implications
by Shek Kam Tse Hui Li
De Gruyter Mouton
ISBN 978-3-11-024004-7 e-ISBN 978-3-11-024009-2 ISSN 1861-4248 Library of Congress Cataloging-in-Publication Data Tse, Shek Kam, 1948⫺ Early child Cantonese : facts and implications / by Shek Kam Tse, Hui Li. p. cm. ⫺ (Studies on language acquisition; 42) Includes bibliographical references and index. ISBN 978-3-11-024004-7 (alk. paper) 1. Cantonese Dialects ⫺ Grammar. 2. Language acquisition. 3. Early childhood education. I. Li, Hui, 1968 Sept. 28⫺ II. Title. PL1733.T74 2011 495.11795127⫺dc23 2011028265
Bibliographic information published by the Deutsche Nationalbibliothek The Deutsche Nationalbibliothek lists this publication in the Deutsche Nationalbibliografie; detailed bibliographic data are available in the Internet at http://dnb.d-nb.de. 쑔 2011 Walter de Gruyter GmbH & Co. KG, Berlin/Boston Printing: Hubert & Co. GmbH & Co. KG, Göttingen ⬁ Printed on acid-free paper Printed in Germany www.degruyter.com
Foreword Cantonese, the third most widely spoken Chinese language dialect, is used by over 70 million people1. The vernacular is spoken in Southern China, Hong Kong, and Macau. In addition, considering that many overseas Chinese from Guangdong (formerly Canton) Province have emigrated, Cantonese is also commonly heard in overseas Chinese communities. However, given that Cantonese has no official standard written form, it is neither taught in schools nor used in government communications. China has experienced tremendous economic growth and is poised to soon become the world’s second largest economy. Investment in research has led to a phenomenal increase in scientific studies and China is expected to lead the world in the production of scientific research by the end of this decade. Although the literature on child development in China is burgeoning, there are few methodologically rigorous large-scale studies of Chinese language acquisition. Given the pervasiveness of Mandarin Chinese, even fewer research projects have been concerned with the acquisition of Cantonese. In light of these situations, a volume on early childhood Cantonese is particularly necessary and timely. The authors of this book, Shek-Kam Tse and Hui Li, have extensive research experience in Chinese language development and education. Based on a large corpus of language data gathered from a representative sample of nearly 500 preschoolers in an authentic situation, they have made a seminal contribution to our understanding of Cantonese language acquisition. Through their detailed and carefully analyzed findings, they have reported on the development of preschool children’s vocabulary, grammar, and use of questions and temporal words. I believe that this well-written and comprehensive volume would be appropriate as a reference book for courses on Chinese language acquisition, Chinese linguistics, and early childhood development and education in Chinese societies. The findings on normative Cantonese language development are also relevant to the work of speech therapists working with Chinese populations. Furthermore, the findings presented in this book will be beneficial to those who are interested in language acquisition from a cross-cultural perspective, particularly to scholars who are interested in 1
Mandarin Chinese is the most widely spoken form of spoken Chinese with about 850 million speakers, followed by Wu Chinese, which has about 90 million speakers.
vi Foreword how children acquire the grammatically simple, tonal Cantonese dialect. This book will also be of interest to psychologists, curriculum designers, textbook writers, educators, and policymakers concerned with development and education during the early years. I hope that this very useful and timely book, Early Child Cantonese: Facts and Implications, will be an impetus for further empirical research on language development in Cantonese-speaking children. Nirmala Rao Professor, Faculty of Education and Associate Dean, Graduate School The University of Hong Kong
Preface
Chinese is often considered very difficult to learn by Western language users. Typically, Western language users are puzzled by the ideographic Chinese characters and Chinese grammar due to the significantly different characteristics of Chinese from some common Western languages such as Spanish, French, and English. This perplexity with Chinese further deepens when the written form of Chinese is taken into consideration—the written form varies across dialects such as Cantonese and Mandarin. Nevertheless, these characteristics have inspired Western scholars to learn and investigate effective approaches to learning Chinese. To acquire Chinese, it is important to grasp the most basic language elements first. Therefore, it is inevitable to explore child language as reflected in the present study. In addition, the time between ages three and six is considered to be an important period of language development. During this period, children rapidly acquire their native language and develop their mental and physical capacity. Notably a delay in language development in the preschool years would result in the postponement of the acquisition of knowledge and communicative skills, making it difficult for follow-up education to remedy such a loss. Hence, language development in children is always an important topic in the field of psycholinguistics. In the present scenario, however, kindergartens and parents often concentrate more on the skills in taking entrance examinations for the next school level. Under the very competitive education system, children’s native language development is not given considerable attention. Worse, the current preschool language curriculum has been designed in a way that assumes that children undergo similar life experiences as adults do. Therefore, the so-called “common words” in many textbooks are actually found by counting the frequencies of words in newspapers and magazines. In other words, these “common words” are linked with events in the adults’ world rather than connected with children’s life experiences. In fact, requesting children to speak, learn, and make use of these words ignores their actual developmental needs and capabilities. We published this book to promote the characteristics and importance of early language development. The research in this book was conducted and analyzed from both the educational and linguistic perspectives to provide an understanding of the basic Cantonese usage of Hong Kong preschoolers
viii Preface aged between three years and five years and 11 months. The research data were collected from carefully sampled local kindergartens by recording the children’s utterances in an authentic classroom setting. The commonality in the data was then statistically analyzed by calculating the proportions of the selected language elements such as word and sentence types and syntactic operations in utterances. We hope that this study can serve as a good resource to researchers and educators in their efforts to develop curricula and teaching materials, which are more realistic and in accordance with children’s daily experiences. We sincerely present this work to all educational, developmental, and cognitive psychologists, psycholinguists, educational researchers, speech therapists, and language teachers, as well as anyone who is interested in child development or the Chinese language. If this book helps them better understand the needs of children, and if it provides them with insights for improving their services or teaching strategies, then it has indeed achieved its purposes. Shek-Kam Tse Professor, Faculty of Education and Director, Centre for Advancement of Chinese Language Learning and Research The University of Hong Kong Hui Li Assistant Professor, Faculty of Education The University of Hong Kong
Acknowledgments
We are very grateful to all our colleagues and friends who have collaborated in the development and production of this publication. At each step, from the initial proposal stages to the final production details, we were greatly blessed for having colleagues who have generously shared their time and talents by giving feedback on our work. First of all, we must especially mention that this project was originally proposed and developed by Dr. Sylvia Opper, who has worked with us at the University of Hong Kong for many years. We wish to present this publication as a tribute to our beloved former colleague. Second, we would like to extend our thanks to Professor Peter Jordens (series editor of SOLA), Dr. Ursula Kleinhenz (in-house editor of the Mouton series Studies in Language Acquisition), and Dr. Julie Miess (senior editor at Mouton de Gruyter). Without their support, patience, and cooperation, this publication would not be possible. Third, we profoundly appreciate the support, assistance, and guidance provided by Dr. Terry Dolan, Ms. S. K. Leung, Ms. W. Y. Chio, Dr. C. Chan, Ms. S. M. Kwong, Mr. W. J. Lau, Ms. L.Y. Fung, Miss Lilian Chau, and Eva Lau. In particular, we are deeply indebted to Miss Jessie Wong for her painstaking work in reviewing, editing, and proofreading the text. We also thank Miss Eileen Wong for her excellent work on the early child Cantonese glossary. Their extensive contributions have made this book a far better work. Fourth, we are very grateful to the participating teachers, principals, and their kindergarteners for their exceptional collaboration and participation in this project. Without their involvement, the corpus of early childhood Cantonese could not be established, and this publication would be impossible. Special thanks must go to Professor Nirmala Rao, associate dean of the Graduate School of the University of Hong Kong. She has been recognized internationally for her excellent work on early childhood development and education. A highly resourceful mentor, Professor Rao wrote the Foreword, which has greatly polished the contents of this book. Finally, we wish to express our sincere thanks to the reviewer of this book. His/her insightful comments and critical review have contributed in numerous positive ways to the overall presentation of this book. Further-
x Acknowledgments more, we take this work as a token of gratitude to all our friends and colleagues who have made the long journey of working at the University of Hong Kong a truly rewarding experience. Shek Kam Tse and Hui Li May 2011
Contents Foreword by Nirmala Rao Preface by Tse and Li Acknowledgments Chapter 1 Introduction 1.1 Cantonese as a major Chinese dialect 1.2 Reasons for studying early childhood Cantonese 1.3 Research data and method 1.4 Organization of the book Chapter 2 Cantonese lexical development in the early years 2.1 Lexical categories in Cantonese 2.2 Early child lexical development: Repertoire, norms, and trends 2.3 Noun bias versus verb bias 2.4 Cantonese classifiers 2.5 Conclusion and implications Chapter 3 Syntactic development in the early years 3.1 Indicators of syntactic development 3.2 Syntactic development in the early years 3.3 Gender and age differences in syntactical development 3.4 Conclusion and implications Chapter 4 Interrogative development in the early years 4.1 Interrogative forms in early childhood Cantonese 4.2 Interrogative functions in early childhood Cantonese 4.3 Conclusion and implications
v vii ix 1 1 6 10 15 22 23 25 31 37 49 50 50 54 62 65 70 70 81 89
xii Contents Chapter 5 92 Cognitive development in the early years: The case of temporal words 5.1 Forms of early Cantonese temporal words 92 5.2 Development of early Cantonese temporal words 96 5.3 Discussion, conclusion, and implications 104 Chapter 6 General conclusion and implications 6.1 General conclusions 6.2 General discussion 6.3 Limitations of the study 6.4 Implications
108 108 109 112 112
Appendix I: Jyutping: The Cantonese Romanization System proposed by the Linguistic Society of Hong Kong
117
Appendix II: Early vocabulary inventory for Cantonese Chinese
119
References
190
Index
201
Chapter 1 Introduction
This opening chapter introduces the background, rationale, and objectives of the studies presented in this book. A brief overview of Cantonese is presented, and the differences between Cantonese and Mandarin are discussed. Summary accounts of early language acquisition theories are outlined, and the purposes of studying young children’s acquisition of Cantonese are likewise explained. Consequently, there is an outline of the main body of research carried out to gather the evidence used in the studies presented in the various chapters. Each chapter reports the research strategy and outcomes, the conclusions generated from each particular study, together with the implications for further research and the development of children’s Cantonese proficiency in the early years of schooling. 1. Cantonese as a major Chinese dialect The Chinese language is the most widely spoken language in the world today (Weber 1997). Given the immense geographical topography of China’s many provinces, several Chinese regional dialects have developed over the years. To achieve a degree of uniformity in language usage, Mandarin (also known as Putonghua) has been designated as the official spoken language across Mainland China. Mandarin is also spoken in Taiwan, and it is one of the official languages of Singapore. Standard Written Chinese (SWC) is the written equivalent of spoken Mandarin, and its script is used across all regions of Mainland China, regardless of the local dialect. Meanwhile, Cantonese, one of the many major Chinese dialects of Chinese, is currently used as the lingua franca in Southern China, Hong Kong, Macau, and many “Chinatowns” in North America. Majority of overseas Chinese communities originate from Southern China; hence, most of them speak Cantonese either as their mother tongue or as a second language. The word “Cantonese” is associated with the word “Canton”, used by early European merchants for the region around the capital city of Guangdong. Although it is spoken by some 70 million people, Cantonese has no official standard writ-
2 Shek Kam Tse and Hui Li ten form and is neither taught in schools nor used in formal business and government communications. Although there are five major sub-dialects of spoken Cantonese (Yue–Hashimoto 1991), this book only focuses on the Cantonese spoken in Hong Kong as the everyday form of verbal communication at home and in commerce, media and schools. 1.1. Phonology of the Cantonese language 1.1.1. Syllables Like Mandarin and other Chinese dialects, Cantonese characters are monosyllabic. However, it is possible for each character to have multiple syllable pronunciations. A syllable may also correspond to multiple character homophones (Tse 2006). Cantonese and Mandarin belong to the family of Sino-Tibetan languages, which are also monosyllabic and tonal. However, there are important linguistic differences between the two, including “the inventory of syllable segments, the tone system and lexical and grammatical structures” (Matthews and Yip 1994). As an example, Cantonese has 625 base syllables (tone-independent monosyllable units), which are over a third more than the 420 syllables in Mandarin (Lee, Lo, Ching, and Meng 2002). 1.1.2. Initials and finals In spoken Mandarin, initials or beginning sounds are always consonants. Finals are the ending sounds, which always begin and may end with vowels, consonants or diphthongs. In the analysis of Lee, Lo, Ching, and Meng (2002), there are 23 initials and 37 finals in Mandarin, whereas 19 initials and 53 finals are found in Cantonese. The number of finals in Mandarin is less than that in Cantonese because the latter “has six different consonant codas, whereas the former only has two” (Lee et al. 2002). Generally, the 53 Cantonese finals can be broadly classified into five categories according to Lee et al. These categories are “simple vowels, diphthongs, vowels with nasal coda, vowels with stop coda, and syllabic nasals”. Lee et al. further stated that, “except for syllabic nasals, all Cantonese syllables contain at least one vowel element”. In Cantonese, the initials are optional and all consonants can be present as initials, but not all initials are conso-
Introduction 3
nants. Some initials are semi-vowels or nasals. Non-nasal initials include liquids, glides, fricatives, affricates, and plosives. Like many languages whose pronunciation of words has changed over time, Hong Kong Cantonese contains evidence of exposure to other world languages such as Japanese and English. “Loanwords” from external languages have gradually crept into Cantonese, and there is usually a measure of diversification in the initials and finals used in Cantonese, even by linguists and other scholars. Table 1.1 (page 16) shows the initials system proposed by the Linguistic Society of Hong Kong (2002) with a complement suggested by Bauer and Paul (1997) on the existence of noncontrastive, palatalized allophones. Table 1.2 shows the vowel system proposed by Bauer and Paul (1997) with reference to the Linguistic Society of Hong Kong’s (2002) Romanization of International Phonetic Alphabet (IPA) in Cantonese. Table 1.2. Inventory of Cantonese finals as represented in IPA Final Consonant Nil -i -u -y -m -n -ŋ -p -t -ik
i iu
u
y
ui
a ai au
œ
ũ ɐ ɐ u
Ţ Ţi
ei
ũu Ɔy
im in iŋ ip it ik
un
ut
yn
yt
am an aŋ ap at ak
ɐ m ɐ n ɐ ŋ ɐ p ɐ t ɐ k
ũm ũn ũŋ ũp ũt ũk
m
Ɔn œŋ œt œk
Ɔt
Ţn Ţŋ
ŋ
Ţt Ţk
For the convenience of presenting utterance information to non-Chinese readers, the transliteration symbols presented in this book use the Jyutping system, which was developed by the Linguistic Society of Hong Kong and is interchangeable with IPA symbols. A cross-reference table from the Linguistic Society of Hong Kong (2002) is shown in extract in Appendix I.
4 Shek Kam Tse and Hui Li 1.1.3. Tones Tones in monosyllabic languages are regarded as a feature of pitch movement within syllables. Cantonese is a tonal language whose every syllable contains usually one lexical tone. The identification of a Cantonese tone relies on both the pitch contour and the relative pitch level. The duration of Cantonese tones also plays an important role in separating the so-called “entering’” tones from others. Tone identification is essential for recognizing Cantonese, which is renowned for its richness in tones. In Mandarin, only five tones are used, whereas in Cantonese, nine different lexical tones are recognized. A six-tone classification of Cantonese spoken in Hong Kong is commonly adopted and agreed upon by local linguists: (1) high-level tone (or high-falling tone), (2) mid-rising tone, (3) mid-level tone, (4) low-falling tone, (5) low-rising tone, and (6) low-level tone. In traditional rhyming, the first and fourth tones are regarded as “flat category”, whereas the other tones are said to be “oblique”. (See Table 1.3 on page 17.) 1.2. Cantonese versus Mandarin Cantonese and Mandarin have both developed from Middle Chinese. Moreover, they both maintain the tone system commonly used in ancient times. Cantonese used to be the main language of southern China before the port-opening policy that commenced in the Qing Dynasty. At the time, Canton (Guangdong) was a relatively stable and cohesive region in terms of political, cultural, economic, and socio-linguistic development. In general, the Cantonese dialect possessed several linguistic characteristics of Middle Chinese, including the structured tonal system. For instance, it featured the three kinds of tone levelling: rising and departing tones with the final “p”, “t”, and “k”. Also, some words in the popular lexicon in Cantonese have been maintained and used as they had been in Middle Chinese, for example, “Ὢ” (hai6: to be) “㗷” (gan3: late), “浣䨩” (wat6 dat6: ugly). These words may be found in classical Chinese dictionaries; however, they are no longer used in Mandarin. In contrast, Northern China has been characterised by a more chequered history, unstable politics, and fluctuating economic development. Its people had frequent contact with other ethnic Chinese groupings and with those who spoke foreign languages. Consequently, although Mandarin is also a branch of Middle Chinese, the phonetic system in the North changed considerably as it fused with other contact
Introduction 5
languages and absorbed words popularly spoken in neighbouring cultural and trading environments. Therefore, although there are lexical differences between Cantonese and Mandarin, and the discourse exchanges in the two linguistic cultural groups use local social and phonetic inventories, the syntax and grammar in the two dialects carry distinct similarities. 1.3. The Romanization of Cantonese There are several Romanization systems presently used for writing phonetic versions of spoken Chinese. In the case of Cantonese, the Yale Romanization system is commonly used in the United States. In 1993, The Linguistic Society of Hong Kong proposed a new standardized system called Jyutping. This new system set out to avoid some of the problems experienced in other Romanization schemes (such as reformulating the tones, consonants, and coda of the Cantonese phonetic inventory) employed as a strategy for achieving consistency and uniformity. The form of Romanization used in the Jyutping system employs the English alphabet and morphemes to achieve consistency and standardization in print. The Jyutping scheme (http://www.lshk.org/cantonese.php) is used throughout this book.
Question Box: 1. What is Cantonese? Who speaks Cantonese in their everyday life? Where do they live? 2. What are the differences between Cantonese and Mandarin? Why and how have differences emerged in the evolution of the two languages? 3. How is Cantonese Romanized? 4. How many initials and finals are there in Cantonese? How many combinations of these are there?
6 Shek Kam Tse and Hui Li 2. Reasons for studying early childhood Cantonese 2.1. Cantonese deserves systematic research 2.1.1. Cantonese is ancient as well as modern As spoken today, Cantonese features certain phonological characteristics that are quite close to some pronunciations in Middle Chinese. In fact, some classic Chinese literary pieces, especially poetry, sound close to the original when read in Cantonese because Cantonese phonology is similar in many aspects to the pronunciations used in Middle Chinese. In contrast, the pronunciation of Mandarin is very different from Middle Chinese; thus, poetry in Middle Chinese can sound quite discordant when read in Mandarin. Using Mandarin phonology in poetry and other rhyme-based writings originally composed in Middle Chinese may sometimes make the lines of the poem sound quite incoherent as Mandarin differs so markedly from Middle Chinese phonology (for a review, see Newman and Raman 1999). Cantonese is generally a “word-of-mouth” language; hence, almost all Cantonese-speaking people speak colloquial Cantonese in their daily interactions. The spoken form of Standard Written Chinese (SWC), also known as “High Cantonese”, is quite formalized and is only used in official communications. As a living, dialectal language, Cantonese contains numerous slang words popularly used in the environment and the media. The influence of mass media and pop culture spans many years; hence, the language of the young people of Hong Kong constantly changes and evolves. Moreover, the impact of English on Cantonese is considerable, as many “loanwords” used in contemporary Cantonese appear to be “borrowed” from English. In fact, many of these loanwords have been transliterated with Chinese characters by people who map the pronunciations of English words to those in Cantonese, for example, baa1 si2 ‘bus’. As a result, new examples of slang, trendy expressions and English words constantly appear in the Cantonese language.
Introduction 7
2.1.2. Cantonese is localized as well as globalised Cantonese is localized because it is mostly spoken by people in Guangdong, Guangxi, the Pearl River Delta, and adjacent regions such as Hong Kong and Macau. However, Cantonese is also to an extent globalised, as it is used internationally by millions of ethnic Chinese speakers around the world. For example, it is spoken by Chinese in many South-East Asian countries, including Thailand, Malaysia, and Singapore. Cantonese is also spoken by immigrants from Hong Kong and the Guangdong area, as well as those in areas (i.e., “Chinatowns”) in foreign cities such as San Francisco, New York, Vancouver, Toronto, and Melbourne. Nevertheless, Mandarin is the officially designated medium of communication and education in Mainland China. It is even prescribed for official functions in Cantonesespeaking regions. 2.1.3. Cantonese is both “dying” and “evolving” As a dialectal language, Cantonese can be viewed as both “dying” and “evolving”. It is arguably dying considering that Mandarin is the official as well as the dominant language in Mainland China. Mandarin has been strongly prescribed in Hong Kong since the former British colony’s return to Chinese sovereignty in 1997. Hence, an outsider might expect the number of people using Cantonese to diminish and the number of people using Mandarin to increase. However, viewed from a sociolinguistic perspective, Cantonese is clearly a developing, vibrant language and can be said to have two written forms in Hong Kong: the formal form and the colloquial version. Formal written Cantonese follows the syntax, semantics, and lexicon of SWC to maintain a standardized written communication with people using other Chinese dialects. Although the written form of colloquial Cantonese adopts standard Chinese characters, localized characters regularly surface in Hong Kong Cantonese alongside the colloquial usage of Cantonese grammar. In fact, special characters appearing in SWC often evolve from other languages such as English and Japanese. Sometimes, they may even be created by Hong Kong people themselves. This applies particularly to the Cantonese spoken by people in non-Chinese locations and in Internet chat rooms. Moreover, a number of local newspapers are now written in “colloquial Cantonese”, which although well received by Hong Kong people, is found to be tiring and difficult to read by many Mandarin speakers who do not know Canto-
8 Shek Kam Tse and Hui Li nese. In other words, although Cantonese may be viewed as “dying” by some, it is in fact growing and evolving linguistically over time. 2.2. Early childhood Cantonese deserves empirical exploration Classic studies of language development from infancy are renowned in the West, all of which are relevant to learners of Chinese. Piaget (1926) studied the development of language and thought in infants in French-speaking Switzerland, whereas Vygotsky (1962) looked at the relationship between language and thought development among children in Russia. Skinner (1957) made a systematic study in the United States of how “language behaviour” is operantly acquired through conditioning like any other behaviour; on the other hand, Chomsky (1965) argued for the presence of a language acquisition device, regarded as an innately inherited capability of all humans. Particularly, he looked into syntax, sentences, and language acquisition, as well as how children seem to “invent” language structures they have not previously encountered. From this nascent research platform, Western research then seemed to take a variety of diverse paths, with the attention focusing on the growth of language proficiency in children in early childhood classrooms (Brown 1973; Carol and Carol 1985). Western studies have focused on psycholinguistics and information processing and on applied language in the form of reading and writing and its pivotal role in education (Lunzer and Gardner 1979). More and more studies have examined language in the classroom, how different approaches to teaching language succeed or fail, and the vital importance of effective language development for scholastic performance (Bullock 1975). With the burgeoning emergence of Chinese language scholars in recent years, more and more studies have been conducted on the development and education of Chinese children. However, little academic research has been reported about the early acquisition of Chinese languages. The impetus to look closely at the language used for learning and teaching in Hong Kong accelerated with the admission and knowledge that Cantonese is the medium of instruction in most Hong Kong kindergartens and primary schools. It is also used in many secondary schools, especially to explain concepts that are difficult to understand when presented in English. Researching Chinese children’s language development swiftly gathered momentum in the last few decades, with studies of how spoken Chinese could best be taught and how ways of writing characters could accelerate children’s learning (Tse 2007). However, the bulk of formal research has
Introduction 9
been directed towards attaining competence in Mandarin. Cantonese and Mandarin differ in their phonologies, vocabulary, and grammatical structures (Li and Shirai 2000); hence, there is a need for a systematic and intensive study of language development in Cantonese-speaking children. In fact, in recent years there has been a gradual shift in Hong Kong towards studying Cantonese. Most published academic studies have tended to focus on primary school and junior high school children’s acquisition of the Chinese language (Chan 1985; Hong Kong Education Department 1986; Tse 1990; Tse and Cheung 1991; Tse and Lam 1992). Furthermore, most of these studies have tended to focus on language channels in isolation, for example, lexical development (Hong Kong Education Department 1986) and writing (Tse 1990; Tse and Cheung 1991; Tse and Lam 1992). Thus, there is still doubt today about the exact number of words in the productive repertoire of young Cantonese-speaking children. Likewise, there is no consensus among researchers about children’s productive repertoire of Cantonese temporal words. For example, Fletcher, Leonard, Stokes, and Wong (2005) suggest that preschoolers can produce at least six aspect makers, whereas Matthew and Yip (1994) and Yip and Matthew (2000, 2001) suggest 13 markers as the baseline. In addition, Cantonese-speaking preschoolers’ productive repertoire of interrogatives has not been extensively and systematically examined in previous studies. In effect, there is a dearth in the number of qualitative and systematic studies of Cantonese acquisition in early childhood and in the first years of formal education. Given the paucity of research studies in the field, there are difficulties associated with the design of preschool syllabi and curriculum design and evaluation. What is needed is a drive to conduct empirical research with representative samples of Hong Kong preschoolers, especially research that systematically looks into their early childhood Cantonese development. Researchers in the present study were inspired to conduct a large-scale study of Cantonese acquisition in the early years to address this literature gap. It is hoped that this publication will make a significant contribution to the field of early childhood Cantonese development, and that it will stimulate awareness of important implications for policy making, teacher education, and curriculum design. These policies may influence the allocation of resources for early childhood Cantonese education and pull together the various strands that do not always move ahead in the same direction.
10 Shek Kam Tse and Hui Li Question Box: 1. Why do we say Cantonese is ancient as well as modern? 2. Why is Cantonese localized as well as globalised? 3. Is Cantonese “dying” or “evolving”? List examples of Cantonese words to support your answer. 4. Why is early childhood language study so important? 5. Why should early childhood Cantonese be studied? What are the gaps in the published literature? 3. Research data and method Investigations in this book are based on analyses of a representative sample of children aged three to six years who are currently receiving preschool education in Hong Kong. Evidence was collected in naturalistic settings in which all of the children’s conversations were recorded and later transcribed. The language corpus was based upon the conversations of 492 randomly selected Cantonese-speaking children as they participated in pairs in a 30-minute play task. The utterances recorded enabled a wide range of linguistic features of the spoken Cantonese of the preschool children to be identified and logged in detail. The corpus of evidence allows researchers and people with access to the data to obtain a detailed overview of developmental features of Cantonese language acquisition among preschool children in Hong Kong. 3.1. Corpus of evidence approach A linguistic corpus typically contains collections of text that represent samples of how the target language is actually used in authentic settings. In its textual form, the language is machine readable and, in addition to being assessed for qualities such as its readability, is available for a range of analytical techniques. Many of these analytical techniques would be very tedious and time-consuming if executed by hand. A language corpus provides material for discourse and statistical analysis, and allows analysts to assemble empirical evidence that may yield developmental trends in early childhood language. There are advantages associated with the use of a corpus approach. First of all, a well-designed and comprehensive corpus offers a database availa-
Introduction 11
ble for multi-purpose analyses by independent researchers. Second, using the corpus as a primary source of evidence allows theories to be developed and/or tested and retested, and the generality of the findings to be tested against the output of studies of similar language databanks. Third, a rich corpus allows researchers to return to original sources to check the validity of tentative or hypothetical analyses. On the other hand, a major limitation of corpus-based research is that the data only represent a sample of the entire language, not the target language in its totality. Any natural corpus is likely to be skewed in its content and pertinence for the wider scenario. Moreover, its users need to keep in mind that no particular corpus will contain information about all of the areas of language that a linguistic researcher may want to explore. To minimize this eventuality, the corpus needs to be comprehensive and inclusive, and should be derived from a large sample with data collected from typical, everyday contexts. Considering the comparative infancy of research into early childhood Cantonese, the non-usage of the experimental approach as a research strategy was decided. It would have been possible to present children with linguistic input and ask them to perform tasks, and then measure performance in terms of differences in reaction times, utterances, and appropriateness of responses. Nevertheless, there are benefits associated with the use of the experimental approach, including the following: the ability of the researcher to control variables and focus on specific aspects of performance; controls can be applied to allow for conclusions to be drawn about specified experimental groups; and the experimenter may modify or manipulate conditions and hold factors constant whilst observing the effect of variations on the experimental subjects. However, such rigour is best applied when all the variables can be controlled and if it can be claimed that generalizations based on the results are valid. Schools and classrooms are not laboratories. The artificial setting of control and experimental groups is not always feasible, and huge samples need to be involved for any result to merit claims of generality. 3.2. Sample The data in this book were extracted from a corpus of language evidence accumulated by Tse and his colleagues in the past decade (Tse, Chan, Kwong and Li 2002; Tse, Chan, and Li 2005; Tse, Li, and Leung 2007). Subjects in the study were randomly sampled as participants in the International Educational Achievement (IEA) Pre-primary Project, aimed at col-
12 Shek Kam Tse and Hui Li lecting data from Hong Kong children to establish a normative framework of developmental milestones for children aged 36 to 71 months. The sampling frame of the IEA Project consisted of all Cantonese-speaking students aged 36 to 71 months who were receiving preschool education in Hong Kong. The final sample in the study consisted of 492 children selected from 68 preschools (58 kindergartens and 10 nurseries) located in Hong Kong Island, Kowloon, and the New Territories. The subjects were randomly selected from each class in the participating preschools, and comprised three age groups (three-, four-, and five-year-olds), with 82 boys and 82 girls in each age group. To minimize the influence of variability in family background and education, the sample of children in each preschool was limited to 10 subjects. Table 1.4 provides details about the regions and participating preschools. Table 1.4. Regions and selected preschools Region Kindergarten
Nursery
Total
Large Scale Small Scale Large Scale Small Scale Hong Kong Island
13
10
2
2
27
Kowloon
11
4
0
3
18
New Territories
12
8
0
3
23
Total
36
22
2
8
68
3.3. Communication task used to gather information and evidence A free-play corner was set up in the classroom of the children selected. The area was furnished with a set of toys that included cooking materials, food, fruit, furniture, electrical appliances, hospital-related materials, and toy vehicles (Table 1.5). Each randomly arranged pair of subjects (boy/girl, boy/boy, or girl/girl) was left in the free-play corner for 30 minutes. They were encouraged to talk whilst playing, and the 30-minute conversations were recorded in both video and audio formats. During the free-play sessions, researchers were allowed to observe and casually oversee the children. However, researchers were prohibited from intervening in the children’s activities. There were no other children in the room.
Introduction 13 Table 1.5. Test materials displayed in the free-play corners Category Item Cooking material
Cooker with oven and rings, a rice cooker, an apron, chopsticks, knife, plates, bowls, pots, pans, and spoons
Food and fruit
Bunch of grapes, assorted vegetables, bananas, bottles of fruit juice, bread and buns, cake, candies, eggs, fish, fork, milk cartons, packet of cookies, roast chicken, two apples, and two oranges
Furniture and electrical Refrigerator, table, telephone, TV set, bed, chairs, and pot appliances of flowers Hospital materials
Recipe/notebook, clinical thermometer, flashlight, stethoscope, and two pencils
Vehicles
Car, plane, ship, and train
The conversations were recorded on tape then transformed into digital format for later analysis. Audio conversations were transcribed into text by researchers and included all audible single words and overlapping conversations. Non-lexical registers and other sounds in speech uttered by children, for instance, exclamations indicated by single syllables in Cantonese, were also transcribed. Independent researchers re-checked all the transcriptions by listening to the records again to ensure the accuracy of the transcriptions. 3.4. Coding of the lexical items Although widely accepted definitions of a ci ‘word’ do not exist in Chinese linguistics, the writers hold the widely accepted view (see Wang and Zou 1998) that a word is the smallest linguistic unit that can express meaning independently. In the present study, words were grouped as different classes according to their properties and the ways they were functioning in the sentences, following a classification system modified from the investigations of Ko (1980) and Cheng (1994). The final database consisted of a total of 90,908 words from the 492 young children. As mentioned earlier, Cantonese is a dialectal language and its syntax and uses of lexical items differ from those in Mandarin. Consequently, the classification of lexical items devised by linguists for the purpose of analysing traditional Chinese may not totally apply. In this regard, the re-
14 Shek Kam Tse and Hui Li searchers adopted the categorization scheme suggested by Cheng (1994) for recoding Cantonese transcriptions to ensure that the coding was exhaustive and truly covered all the words in the conversations (Table 1.6, pages 18–21). 3.5. Coding of sentences Sentence development in children is usually measured by the average length of uttered sentences, and sentence structure and complexity. The analysis of sentence structures and complexity of utterances included in the research frameworks was performed by various scholars (Chao 1968; Crystal, Fletcher, and Garman 1976; Kwong 1990; Li 1968; Miller and Chapman 1983; Scarborough 1985). Narratives, interrogatives, imperatives, negations, exclamations, and incomplete sentences were the major types of sentence scrutinised in the present research, with particular considerations as follows: –
– – – –
Each child’s oral language was segmented into units of “utterance”. Lund and Duchan’s (1988) principles on judgments of utterance boundaries were employed, with certain modifications to accommodate the characteristics of each utterance made by the children; Generally, an utterance ended by a slump or a boost in phonological inclination; Although a speaker may have spoken elements from several sentences at the same time in a jumbled manner, each sentence was treated as a separate unit of analysis; Two sentences with both “subject-predicate” structures joined by a connective (compound sentence) were counted as one analysis unit; In the Chinese language, a compound sentence is not necessarily connected by conjunctions. The semantic relationship between two sentences depends on the structure of the two subjectpredicate constructions. This type of sentence is also counted as a single unit of analysis.
Introduction 15
Question Box: 1. What is the scale of the language corpus gathered for use in the current study? How many children were tested? Does this corpus represent all of the Cantonese-speaking children in Hong Kong? 2. What is your opinion about the communication task? Was the communication task adequate for assembling data about the children’s productive speech? 3. Are there other limitations in the current study? Share your ideas with your colleagues.
4. Organization of the book This book is organized into six chapters. Chapter 1 introduces the background, purpose, approaches, and significance of the studies presented in this volume. Chapter 2 reports the findings about the lexical development in Cantonese-speaking young children, with a focus on the early Cantonese lexicon, the “noun-bias” versus the “verb-bias” debate, and on classifiers. Chapter 3 focuses on syntactic development in the early years, introducing the Cantonese syntax system and developmental trends. Chapter 4 presents findings about the forms and functions of early Cantonese interrogatives. Chapter 5 discusses the relationship between cognitive development and language development in light of the findings about early child temporal words and time expressions. Chapter 6 draws together the general findings and discusses the limitations and implications of the studies presented in this book. After each section within the chapters, there are questions to help the reader systematically review the major points and findings presented in the text.
p t h h t p Ƒ*
f s Ƶ* ts
Ƶh* tsh
m n Ǧ l
j ȗ
Ǧ
ǡȋͳͻͻȌǤ
Table1.1. Ǧ Ǧ k kh ă w
kw kwh
Ǧ h
16 Shek Kam Tse and Hui Li
Table1.3. Ǧ Ǧ Ǧ ͳ ʹ ͵ Ͷ ͷ ၃ ў ၂ ਔ ѱ
Ǧǡ Ǧ Ǧ Ǧ Ǧ Ǧ ǡ sÄ,sì sí si sÄh,sìh síh sih
Ǧ
ࢂ
Ǧ
Introduction 17
Verb
Temporal
Directional
Imperative
tivity Existence
Ӹӧ ‘exist’, Ѩ ‘disappear’, วғ ‘happen’, 㥀 ‘present’ ٬ ‘let’, ڮз ‘order’, ा‘ request’ ѐ ‘go to’, ٰ ‘come’ ໒‘ ۈstart’, ่‘ ״end’
ྰη ‘lion’, Դ߁ ‘tiger’ Concrete entities ፐҁ ‘textbook’, ᐋЕ ‘tree’, ႝ၉ ‘telephone’ Conceptual entities ௲‘ ػeducation’, Б‘ ݤmethod’, ୢᚒ ‘problem’ Places ३ෝ ‘Hong Kong’, ٥ࢪ ‘Asia’, Ꮲਠ ‘school’ Temporal units ដᓐ ‘hour’, Д ‘month’ Position Εय़ ‘inside somewhere’, ኴ ‘upstairs’, ኴΠ ‘downstairs’ Action າ ‘run’, ઢ ‘laugh’ Development ғߏ ‘grow’, ቚу ‘increase’, ᘉε ‘enlarge’ Psychological ac-׆ఈ ‘hope’, ᇡࣁ ‘deem’, ‘know’, ⮄ ‘think’
Animal
Table 1.6. Classification of Cantonese lexical items (Cheng 1994) Category Description of Examples Category Divisions Noun People Դৣ ‘teacher’, ᙴғ ‘doctor’, ‘ ݿݿpa-pa’
18 Shek Kam Tse and Hui Li
Pronoun
Adverb
Adjective
Person
Other
Mood
Range
Affirmative
Frequency
Magnitude
Judgmental
Modal
ͳǤ ǣךǮ ȀǯǡգǮǯǡᥝǮȀȀȀǯ ʹǤ ǣך䘂ǮȀǯǡգ䘂Ǯǯǡᥝ䘂ǮȀǯ
ӳ ‘very/a lot of’, ϼ ‘too much/too many’, Կ ‘most’, ளᅉ ‘too much/too many’ Ξ ‘again’, ӆ ‘again’, ۳۳ ‘usually’, ԋВ ‘always’ Ѹฅ ‘must’, দ ‘not’, ᢲ ‘nil’, দӳ ‘don’t’, ࠓ ‘don’t’, দᎭ ‘don’t need to’ ӄ ‘all’, ‘also’, ਘ ‘all’, ృ߯ ‘only’ দ೯ ‘isn’t it’, ᙁ‘ ޔjust as’, Կ ‘only/just’, ᐉு ‘anyway’ 䘄䘄ӳ ‘just right’, ݀ฅ ‘just as expected’, ँฅ ‘suddenly’, ۩Զ ‘fortunately’
ૈ ‘can’, ᔈ၀ ‘ought to’, ޭ ‘to be willing to’ ߯ ‘to be’, দ߯ ‘not to be’, ՟ ‘similar to’ এ ‘cold’, ӭ ‘many’, ‘ זfast’
Introduction 19
Classifier
Auxiliary Word
Verb
Noun
Aspect
Mood
Structural
Locative
Demonstrative
Interrogative
ঁ (classifier of concrete objects), చ (classifier of long and thin objects) ԛ (classifier of “attempts”)
ple meanings derived from the context, syntactic (e.g., sentence structure), and suprasegmental features (e.g., inclination) and cannot be simply generalized in a “wordper-meaning” approach. ֹ ‘an action has just finished’, ၸ ‘an action was done some time ago’, ᆙ ‘an action is undergoing’
ࡋګǮǯǡ䛧ࡋ Ǯǯ 㽪 ‘possess’ 㽦 (showing affirmative), (showing imperative). Most of the particles have multi-
ǣᗺǮǯ ǣᢥ䘇Ǯǯ ǣ൳ӭǮȀ
ǯ
ǣᜐࡋǮǯ ǣ൳ਔǮǯǡ൳ऐǮȋȌǯ ǣᜐঁǮǯ ‘ ঁګthis one’, 䛧ঁ ‘that one’
ͳǤ ʹǤ ͵Ǥ ͶǤ ͷǤ Ǥ
20 Shek Kam Tse and Hui Li
Acknowledging Words
Connectives
Locative
Adducted after a noun or pronoun to show locations, e.g., ኴ ‘upstairs’, ኴΠ ‘downstairs’ a. ӕȀӕশȀᆶȀϷȀሸ Ǯǯǡ܈Ȁޣ܈Ǯǯ b. ᩙǮǯ c. ՠ߯Ǯǯ d. ฅࡕǡၟՐǮǯ e. җܭǮ
ǯǡѝाǮǯǡᗨฅǮǯǡԶЪǮǯǡӢࣁ Ǯ
ǯǡ܌аǮǯ ջ߯ Ǯǯǡӵ Ǯǯ
ԭ ‘hundred’, ί ‘thousand’, ‘ten-thousand’ Use ಃ in front of numeric, e.g., ಃ ‘the first’
Quantifier
Ordinal
䡗 (used when showing things to others), ቐ ‘wow’, ࠋ֔ ‘oh’, ༕ ‘yes’, ভ ‘yup’
Exclamation
Cardinal
䛦 ‘at’, Ծ/வ ‘from’, ӛ ‘towards’
Preposition
Introduction 21
Chapter 2 Cantonese lexical development in the early years
Linguistic studies of early lexical development in the West have tended to focus on the patterns of vocabulary acquisition in young children, utterance content, types and categories of syntactic and semantic usage, and the developmental repertories of the lexicons of infants and young children. Attention is also directed to the acquisition of tenses, plurals, and other aspects of grammar by young children. In addition, developmental and cognitive psychologists in the West have looked in detail at the relationship between children’s language growth and their cognitive development. The structure of the lexicon has been investigated from the perspective of young children’s grasp of semantic and syntactic aspects of language, and by psycholinguists interested in the lexicalization of young children’s understanding of the world around them. From a different standpoint, words belonging to a variety of word classes have been found to play different roles in the lexical organisation of different languages (Lehman 1990). This idea bears important implications for the analysis of first and second language acquisition. This observation is important for the study of languages that do not change words to indicate plurality, tenses, and so on, such as the Chinese language. Jiang (2004) and other researchers have tentatively explored the relevance of Western theories for the learning of Chinese. Liang (2004) warns educators that the Chinese language has characteristics that invalidate blanket generalizations from Western language theorizing to studies of Chinese. Nevertheless, like the studies of lexical development in the West, the acquisition of a Cantonese language lexicon involves the construction of four components of words: a phonological component, which involves the ability to identify Cantonese words in a continuous phonetic chain or string; a semantic component, which involves the ability to assign Cantonese words a meaning (content or functional); a syntactic component, which involves socially acceptable grammatical connections between new and known words with particular reference to word classes and to the argument structure of predicates; and a morphological component, which involves aligning new and known words with other words of the same family.
Lexical development 23
This chapter reports the outcomes of an investigation of the lexical development of Cantonese-speaking preschoolers, with focus on repertories and categories in the early body of the Cantonese language uttered in a child-play setting by the Hong Kong children who comprise the sample. Specifically, Chapter 2 sets out the Cantonese lexical classification employed in the studies, reports the repertoire, norms, and trends in the early Cantonese lexicon of the children, addresses the “noun bias” versus “verb bias” debate in light of the outcomes of the present study, examines the emergence of early child classifier usage, and discusses the implications for education in the early stages of learning. 1. Lexical categories in Cantonese Ten lexical categories were elicited from the corpus of utterances recorded in the free-play setting: nouns, verbs, adjectives, adverbs, pronouns, auxiliaries, numerals, classifiers, conjunctions, and prepositions. Noun refers to a word (other than a pronoun) used to identify any class of people, places, or things (common noun), or to name a particular proper noun. In Cantonese, nouns semantically denote objects, things, persons, ideas, or concepts. Syntactically, they are the subject of a sentence or the object, such as lou5 si1 ‘teacher’ and fo2 ce1 ‘train’. Verb refers to a word used to describe an action, a state, or an occurrence forming the main part of the predicate of a sentence. In Cantonese, verbs semantically denote action and activity, and syntactically form the predicate of sentences together with other elements. Examples are mei4 siu3 ‘smile’ and cai3 ‘construct’. Adjective refers to a word naming an attribute of a noun or a noun phrase. Semantically, in Cantonese, an adjective is a word that describes qualities. Syntactically, it is used either to modify nouns or is combined with a verb to form the predicate of a sentence, for instance, dung3 ‘cold’ and mei5 ‘beautiful’. Adverb refers to a word or phrase that typically modifies the meaning of an adjective, a verb, or another adverb, typically expressing manner, place, time, or degree (e.g., gently, here, now, very). Some adverbs, e.g., sentence adverbs, can also be used to modify whole sentences. In Cantonese, the syntactical function of an adverb is to modify a verb, an adjective, or another adverb, such as cyun4 bou6 ‘all’ and ne1 paai4 ‘recently’.
24 Shek Kam Tse and Hui Li Pronoun refers to a word that can function as a noun phrase used by itself and that refers either to the participants in the discourse (e.g., I, you) or to someone or something mentioned elsewhere in the discourse. In Cantonese, a pronoun’s syntactical function is as a substitute for a noun or a noun phrase, such as ngo5 ‘I, me’ and nei5 ‘you’. Cantonese uses pronouns, which apply the same lexicon to function as both subjective (English: I, he, we) and objective (me, him, us) just like in many other Sinitic languages. (See Table 2.1) Table 2.1. Cantonese pronouns Pronoun Romanization Grammatical Classification English Equivalent
ך գ ᥝ ך䘂 գ䘂 ᥝ䘂
ngo5
1st person singular
I/me
nei5
2nd person singular
you
keoi5
3rd person singular
she/he
ngo5 dei6
1st person plural
we/us
nei5 dei6
2nd person plural
you (all)
keoi5 dei6
3rd person plural
they/them
Auxiliary refers to an auxiliary word. Syntactically, in Cantonese, an auxiliary is used to assist a verb to express grammatical function such as tense, aspect, and mood. Examples are gwo3 ‘already’ and gan2 ‘doing’, which are aspect markers widely used in Cantonese. Numeral refers to a word expressing a number. In Cantonese, numerals refer to the lexical class containing Chinese numbers and are used to denote the quantity of objects, things, persons, ideas or concepts. Examples are jat1 ‘one’ and ji6 ‘two’. Classifier refers to a word that indicates the semantic class to which a word belongs. In Cantonese, there are two types of classifiers: noun classifiers measure or classify a noun, whereas syntactically, it appears in-between the noun and the numeral or between the noun and the demonstrative pronoun. Examples are faai3 ‘piece’ and shuang1 ‘pair’. Verb classifiers enumerate the number of times an action has taken place such as da2 saam1 kyun4 ‘hit thrice with the fist’. Connective refers to a word whose function is to link together other linguistic units. In Cantonese, connectives include conjunctions, which are uninflected words used to link together words or sentence elements, and words used to link up phrases, for example waak6 ze2 ‘or’ and tung4 ‘and’.
Lexical development 25
Preposition refers to a word governing (and usually preceding) a noun or pronoun and expressing a relation to another word or element in the clause. In Cantonese, a preposition is a part of speech that is usually indeclinable in form and is used together with a noun phrase to show the relationship between that phrase and the other words in the sentence. Examples are ceoi4 zo2 ‘except’ and hoeng3 zyu6 ‘towards’.
Question Box: 1. How many lexical categories are reported in this chapter? Make a list and give examples for each category. 2. Are there any changes in the form when Cantonese pronouns function as subjective (English: I, he, we) and objective (me, him, us)? Why or why not?
2. Early childhood lexical development: Repertoire, norms, and trends The fieldwork evidence used to illuminate this study revealed that the majority of lexical types voiced by the children were verbs, auxiliaries, pronouns, nouns, and adverbs, accounting for 80% of the total number of words. The most frequently mentioned lexical items were (in order of greatest usage) verbs, auxiliaries, pronouns, nouns, adverbs, classifiers, adjectives, prepositions, numerals, and conjunctions. This ranking was identical across the three age groups, as can be seen in Tables 2.2 and 2.3. The average number of lexical types for each age group was calculated with the total number of lexical types of each group divided by the number of children in the group. Four lexical classes (verb, noun, pronoun, and auxiliary) were featured consistently in the utterances of both boys and girls. Although the distributions of boys’ and girls’ usage of different lexical classes are similar, girls produced relatively more words than boys in all lexical classes analysed in the study (Figure 2.1). In the sections that follow, comment is offered about developments in Cantonese verbs, nouns, adverbs, classifiers, and adjectives spoken by the children in the study.
26 Shek Kam Tse and Hui Li Table 2.2. Averages of lexical types and standard deviations produced by each age group Age 3 Age 4 Age 5 (n = 164) (n = 164) (n = 164) Average
Standard deviation
Average
Standard deviation
Average
Standard deviation
Verb Auxiliary
40.30 29.54
21.79 15.99
41.88 30.43
20.00 16.26
47.17 33.13
25.40 18.54
Pronoun
28.70
15.99
28.87
14.94
29.58
18.13
Noun
23.77
15.46
22.14
12.60
24.79
15.44
Adverb
14.46
9.90
17.02
10.04
17.62
11.54
Classifier
11.07
7.42
12.08
7.41
13.07
9.85
Adjective
8.41
6.20
8014
5.39
9.70
6.86
Preposition
2.27
3.11
2.12
3.13
1.99
2.48
Numeral
1.98
3.02
4.78
10.85
4.66
14.09
Conjunction 0.64
1.18
1.25
5.46
0.72
1.30
Table 2.3. Frequencies of lexical items across the age groups Lexical Class Age 3 Age 4 Age 5
Total
Verb Auxiliary
6,609 5,305
6,869 5,448
7,736 5,964
21,214 16,715
Pronoun
4,932
4,938
5,035
14,905
Noun
3,898
3,631
4,066
11,595
Adverb
2,828
3,340
3,524
9,692
Classifier
1,812
1,981
2,143
5,936
Adjective
1,379
1,353
1,591
4,305
Numeral
324
784
764
1,872
Proposition
373
348
327
1,048
Conjunction
105
205
118
428
Lexical development 27
Figure 2.1. Frequencies of boys’ and girls’ production of each lexical class
2.1. Verbs Verbs were found to be the most frequently observed category. In the present study, a total of 21,214 verbs were uttered during the free-play sessions. Three-year-olds had an average number of 40.3 verb types, whereas four-year-olds had a slightly higher average figure of 41.88 verbs. The average number of verbs used soared to 47.17 in the five-year-old group (Table 2.2). Although the average number of verb types throughout the study increased as children grew older, the proportion of verb usage among all analyzed lexical classes was consistent across the three age groups. The children in the present study utilized verbs to express requests and activities. Sik6 ‘eat’, waan4 ‘play’, zou6 ‘do’, and zyu2 ‘cook’ were frequently spoken in the conversations, and most of the verbs produced by the children were related to dining, as many children engaged in kitchen games in the 30-minute free-play sessions. Although the room was also furnished with furniture and electrical appliances, hospital materials, and toy vehicles, the children invariably acted out cooking/eating sequences. The children sometimes used verbs to construct a noun phrase, for example, saying sik6 faan6 di1 je5 ‘the one needed for eating rice’ instead of saying wun2 ‘bowl’.
28 Shek Kam Tse and Hui Li 2.2. Nouns Nouns appeared to be the fourth most frequently used lexical item, with a total of 11,595 nouns logged during the observation. On average, children aged three, four, and five years spoke 23.77 nouns, 22.14 nouns, and 24.79 nouns, respectively, during the free-play sessions. The proportion of nouns used was consistently lower than that of verbs. The children uttered nouns relating to people and objects that were readily accessible in their daily life such as names of family members and commodities. As the study was set in a free-play environment and most participating children involved themselves in the play kitchen, the majority of nouns they used were nouns about food. In particular, the words faan6 ‘rice’, seoi2 ‘water’, caa4 ‘tea’, daan6 ‘egg’, and min6 baau1 ‘bread’ were often spoken by the children. In the current study, jan4 ‘person’ was the noun most frequently used by the children to generalize persons whose names or associated titles were not known. For indicating locality, uk1 kei1 ‘home’ was the noun most frequently used. 2.3. Adverbs Adverbs were the fifth most frequently used lexical item, and children spoke a total of 9,692 adverbs during the play session. On average, the total number of adverbs spoken by three-, four-, and five-year-old children was 14.46, 17.02, and 17.62, respectively. 2.4. Classifiers The sixth most frequently observed lexical category was the classifier, with a total of 5,936 classifiers identified in the present study. The average number of classifiers used appeared to increase as children grew older (Table 2.2). Table 2.4 shows that the children used a large repertoire of classifiers. The findings indicate that (a) the three-year-olds were able to produce both noun and verb classifiers in the category; (b) more than half (54%) of the noun classifiers were observed in the three-year-olds’ utterances; and (c) another 25% of the noun classifiers were found in the utterances of the four-year olds. The remaining 21% of the noun classifiers were only apparent in the utterances of the five-year-olds. The acquisition and use of Can-
Lexical development 29
tonese classifiers by preschoolers in Hong Kong are discussed in more detail later in this chapter. 2.5. Adjectives A total of 4,305 adjectives were spoken by the children across the three age groups, and the average number of adjectives produced by each three-, four-, and five-year-old was 8.41, 8.14, and 9.70, respectively. The proportion of adjectives spoken was consistent across the three age groups. Table 2.4. Classifiers spoken by the preschoolers (sequencing by quantity of use) Noun Classifiers 1
ঁ
go3
24
Տ
wai2
47
В
jat6
2
ଫ
zek3
25
ҁ
bun2
48
Д
jyut6
3
ࢎ
gaa3
26
ҽ
fan6
49
Չ
hong4
4
݆
bui1
27
bou6
50
tou3
5
ע
baa2
28
ྃ
seoi3
51
ci4
6
చ
tiu4
29
ᅷ
dip6
52
ᛨ
ci4gang1
7
໔
gaan1
30
ဦ
hou6
53
డ
hou4
8
ᙑ
gau6
31
х
bau1
54
൙
wu4
9
༧
faai3
32
౯
hap6
55
฿
tung4
10
zoeng1
33
ᡞ
gun3
56
ᗺ
dim2
11
ಈ
lap1
34
ץ
pai1
57
ᛨ
gang1
12
࿙
wun2
35
௨
paai4
58
ә
cyun3
13
ୡ
daam6
36
᠄
dip6
59
ಔ
zou2
14
ࡋ
dou6
37
ᐅ
zeon1
60
ී
fun2
15
ҹ
gin6
38
ഗ
ding2
61
ࡂ
uk1
16
ኬ
joeng6
39
ᅿ
zung2
62
य़
min6
17
䛥
di1
40
ቫ
cang4
63
gaak4
18
݄
zi1
41
Ԛ
doe2
64
ఀ
zau3
19
૨
man1
42
ᓓ
can1
65
ᒲ
cin4
20
ჹ
deoi3
43
baan1
66
ᆆ
gei1
30 Shek Kam Tse and Hui Li 21
ස
fo2
44
ࣧ
pun4
67
᠈
lung4
22
Ѝ
zi1
45
२
sau2
68
ጃ
seong1
23
൯
fuk1
46
▸
bou1 5
ngaan5
Verb Classifiers 1
ԛ
ci4
3
ৼ
kyun4
2
፺
leong4
4
bou6
Hou2 ‘good’ was the most frequently used adjective, followed by do1 ‘many’ and daai6 ‘big’. Hou2 was usually added before a noun such as hou2 jan4 ‘good person’. Generally, younger children were more likely to use simple adjectives related to things that they themselves had encountered or seen, for example do1 ‘many’, daai6 ‘big’, siu2 ‘small’, and leng3 ‘pretty’. They also used adjectives for temperature and feelings, such as jiy6 ‘hot’ and tung3 ‘pain’. As they grew older, the children were also more likely to use adjectives for describing actions (e.g., faai3 ‘fast’) and physical appearance (e.g., gou1 ‘tall’). Adjectives for describing the state of an event (e.g., lyun6 ‘messy’, co3 ‘wrong’, naan6 ‘difficult’, and gon1 zing6 ‘clean’) and adjectives that are relatively abstract (e.g., zan1 ‘real’, sat6 ‘not hollow’, and gwaai1 ‘well-behaved’) were observed more in the utterances of the older children. Table 2.5 shows the frequencies of several adjectives produced by the participating children. Table 2.5. Frequencies of selected adjectives produced by children three age groups Lexicon Age 3 Age 4 Age 5 hou2 ‘good’ 109 104 120 daai6 ‘big’ 55 70 69 do1 ‘many’ 51 43 44
in the
dak1 ji3 ‘cute’
7
7
19
33
wu1 zou1 ‘dirty’
2
4
6
12
lyun6 xou1 zou1 ‘chaotic’
0
2
0
2
ng5 ngaan4 luk6 sik ‘colourful’
0
0
1
1
Total 333 194 138
Lexical development 31
Question Box: 1. How many lexical categories do early childhood Cantonese children possess? Which category was observed most frequently? Are there any sociolinguistic explanations for this verb-dominating phenomenon? 2. How many types of classifiers can Cantonese preschoolers produce? 3. Which word was the most frequently used adjective? Why? 3. Noun bias versus verb bias 3.1. The debate For many years, studies of both English- and non-English-speaking children have found a predominance of nouns in early utterances and studies of vocabulary acquisition (Gentner 1982; Goldfield and Reznick 1990; Halliday 1975; Markman 1989; McShane 1980; Nelson 1973). The phenomenon of learning nouns earlier than verbs and the nouns dominating children’s early utterances is widely referred to as “noun bias” (Gentner 1982). Despite the dominance of the noun bias assertion, there is anecdotal evidence that the claim may not apply universally, even in the acquisition of English. For example, a longitudinal study by Gopnik (1981) indicated that children generally acquire non-nominal words before nominals, and they tend to use non-nominal expressions in their earliest utterances. In other words, although it could still be true that many English-speaking children learn more nouns than verbs early in their language acquisition, this phenomenon may not apply to all children. In addition, studies of the early acquisition of East Asian languages have yielded evidence that opposes the generality of the noun bias claim. Evidence from studies of Mandarin-speaking children (Tardif, Gelman, and Xu 1999) and young children in Korea (Choi and Gopnik 1995; Kim et al. 2000) has cast doubt on the universality of this phenomenon, with the discovery that verbs appear a lot more frequently than nouns in very young children’s utterances. Choi and Gopnik (1995) investigated Korean-speaking children’s lexical development. They found that most of the children in the study displayed a period of rapid vocabulary growth characterized by a “verb spurt” rather than the “noun spurt” as mentioned by Goldfield and Reznick (1990). Tardif (1996) noted that children “who displayed both a verb spurt and a noun spurt” were characterized by a preceding rapid increase in acquisition of verbs before the rapid increase in the acquisition of nouns. Kim et al. (2000)
32 Shek Kam Tse and Hui Li studied the early productive vocabulary of Korean-learning and Englishlearning children and found that the Korean children learned significantly more verbs than did the English-learning children. Kim et al. (2000) explained this by the more frequent input of socio-pragmatic characteristics by the children; for example, “Korean-speaking caregivers presented more activity-oriented utterances, more verbs, and more salient cues to verbs than did English-speaking caregivers.” Tardif (1996) found that 90% of 22-month-old Mandarin-speaking infants “produced more verbs or action words than nouns or object labels in their natural speech.” Tardif, Gelman, and Xu (1999) later compared the proportions of nouns and verbs in the vocabulary of Mandarin-speaking and English-speaking children and their mothers. They found that Mandarin-speaking children used relatively fewer nouns and more verbs than their English-speaking counterparts; thus, they suggested that context may play an important role in determining the proportions of nouns in the children’s early vocabulary. The authors also argued that whether or not young children’s language acquisition displays noun bias is dependent on many factors, including the method of sampling the vocabulary and the context in which the child’s speech is uttered. The present study found that the Cantonese-speaking children seemed to use more verbs than nouns in their utterances, with nouns ranking low among the lexical categories in terms of incidence and percentages. (See Tables 2.2 and 2.3) To explore whether there is a statistically significant predominance of noun or verb types in Cantonese-speaking children’s vocabulary, the ratio of nouns to verbs was calculated: nouns/(nouns + verbs). The ratio also helps to make an explicit comparison of the data with those of other studies. If the ratio is above 0.50, this indicates that more nouns were produced. In contrast, a ratio below 0.50 indicates that more verbs were produced. As can be seen in Table 2.6, the proportion of nouns to nouns + verbs is approximately 0.37 for the three age groups, which is very close to the ratio (0.38) of Mandarin-speaking children reported by Tardif, Gelman, and Xu (1999). Therefore, all the Cantonese-speaking preschoolers in the present study produced significantly more verbs than nouns in the play context. In addition, Cantonese-speaking and Mandarin-speaking children exhibit a clear preponderance of verbs over nouns in their spontaneous vocabulary as shown in the results of the present study and those of Tardif, Gelman, and Xu (1999).
Lexical development 33 Table 2.6. Comparisons of the noun and verb types and ratios with other relevant data Age Nouns (mean) Verbs (mean) Nouns/(Nouns + Verbs) 36 months (n = 164)
18.76
31.88
0.37
46 months (n = 164)
20.07
33.79
0.37
60 months (n = 164)
22.46
37.66
0.37
A further comparison between noun and verb usage was conducted using the top five nouns and verbs in the frequency lists of vocabulary items used by the children. As presented in Figures 2.2 and 2.3, the frequency of the five most frequently used verbs such as sik6 ‘eat’, waan4 ‘play’, zou6 ‘do’, tai2 ‘look’, and bei2 ‘give’ ranged between 55 and 109, whereas that of the five most frequently used nouns such as jan4 ‘people’, faan6 ‘rice’, seoi2 ‘water’, sin1 saang1 ‘teacher’, and caa4 ‘tea’ varied from 22 to 59. This difference between the frequency of verb and noun types indicates that verbs featured more frequently in the children’s utterances than nouns.
Figure 2.2. The top five verbs produced by the participating children
34 Shek Kam Tse and Hui Li
Figure 2.3. The top five nouns produced by the participating children
The following section addresses the noun-bias versus verb-bias debate, taking into consideration linguistic, cultural, and contextual influences. 3.2. Linguistic, cultural, and contextual influences 3.2.1. Linguistic influences Structurally, Cantonese, like Mandarin, is a pro-drop language, in which verbs are obligatory and rarely omitted (Tardif et al. 1997). For example, Tse (1993) found that, among the 18 types of syntactic structures in Cantonese, 16 require verbs. In contrast, English is a non-pro-drop language, in which more noun phrases, and presumably more common nouns, are needed for communication. In more formal English, overt subject and object noun phrases are usually obligatory, and nouns may occupy a perceptually salient position at the end of a sentence. As a result, a young child’s sentence in English like “Me apple, banana no nice” has no verbs; however, these are implicit, and the meaning is quite clear to English speakers. Furthermore, even regular English verbs may change in sound during declension, and irregular verbs can be a nightmare for the learning of English by non-native speakers. In contrast, verbs are more constant and often occur
Lexical development 35
alone as a complete sentence in Cantonese (Tse 1993). For example, the most frequently used verb by the study’s sample, the word sik6 ‘eat’, can serve as a sentence by itself. In the 30-minute free-play session, when a boy presented an apple to his playmate and said sik6 ‘eat’, the children understood who should eat and what should be eaten on the basis of such a verb sentence. In fact, during the data-gathering sessions, it was common for the children to use verbs to replace nouns temporarily. They appeared to prefer to use phrases including verbs (such as sik6 faan6 di1 je5 ‘the thing that we [use to] eat’) rather than the noun involved wun2 (‘bowl’). In other words, certain linguistic features of Cantonese might help account for the finding that verbs may be more conspicuous than nouns in young children’s Cantonese utterances. 3.2.2. Cultural influences Socio-cultural distinctions between English and Chinese speakers may also have contributed to the apparent differences in linguistic patterns. Choi and Gopnik (1993, 1995) suggested that “English-speaking mothers engage in naming activities with their children more than Korean-speaking mothers do”, and they display a pattern of greater noun than verb usage. In contrast, Korean-speaking and Mandarin-speaking caregivers consistently produced a relatively higher proportion of verb types in their speech vocabulary than did English caregivers (Choi and Gopnik 1993, 1995; Tardif 1993, 1996; Tardif, Gelman, and Xu 1999). In addition, Tardif (1993) found that Mandarin-speaking mothers placed “verbs at the beginnings and ends of utterances with much higher frequencies than they placed nouns”. Although the syntactical and semantic forms of Mandarin used by mothers of young children tend to emphasize verbs, mothers tend to highlight nouns more in the case of English. No speech samples of the Cantonese-speaking mothers were gathered in the present study. Nevertheless, it is reasonable to assume that the type of language children regularly hear around them will influence their daily speech and eventually impact on the predominance of verbs in the early lexicon of Cantonese infants in Hong Kong. Choi and Gopnik (1993, 1995) suggested that “there is a complex bi-directional interaction between the child’s pre-linguistic cognitive structures and the shape of the specific language the child hears in the early stages of language acquisition”. The predominance of verbs in the utterances of adult Cantonese-speakers may have long-term influences on the children’s lexical development. This might
36 Shek Kam Tse and Hui Li help explain why children who are encouraged to focus on verbs in the everyday Cantonese they hear around them acquire verbs early. Thus, it would be inappropriate to claim that the Cantonese-speaking children in the present study produced more verbs simply because of the nature of Cantonese grammar. 3.3.3. Contextual influences In line with the findings of Tardif, Gelman, and Xu (1999), the present study also found that contextual features in the toy play context might have impacted on the subjects’ utterances. Each pair of participating children (boy/boy, girl/girl, or boy/girl) was placed in a free-play corner containing cooking, eating, transport, hospital, and general play artefacts. They were encouraged to talk while they were playing and, not surprisingly, the utterances produced were dominated by verbs associated with cooking, eating, transport, hospitals, and play. Examples of verbs are sik6 ‘eat’, waan4 ‘play’, zou6 ‘do’, tai2 ‘look’, and bei2 ‘give’. As there were just two participants in each session, nouns (subjects) were usually omitted, whereas verbs (actions) were essential in the simple, often incomplete and contextdependent speech. In effect, the free-play context allowed the children to omit nouns when communicating in exactly the same manner favoured by the adult speakers around them. Moreover, the paired children tended to use pronouns, not nouns, to refer to concrete items. Considering that the conversations involved two playmates in a small corner, the children had no difficulty in understanding what the pronouns referred to; for example, ‘ ঁګthis one’, 䛧ঁ ‘that one’, ‘ ࡋګhere’, and 䛧ࡋ ‘there’. Thus, pronouns ranked higher than nouns in terms of frequency of appearance in the present study. Although the fact that context plays a significant role in the speech of Cantonese-speaking children is consistent with the findings of Mandarin-speaking children reported by Tardif, Gelman, and Xu (1999), the children in the present study ranged in age from 36 to 60 months; hence, they could be expected to have the ability to use pronouns freely in their daily speech. Tardif, Gelman and Xu’s (1999) 22-month-olds had just started to use pronouns, and the age difference between the samples may account for the superior use of pronouns by the children in the present study. Although it would seem reasonable to claim that there is evidence opposing the predominance of nouns in the Cantonese-speaking young children’s utterances, this does not imply that steps should be taken by educators
Lexical development 37
to redress an imperfection. Moreover, it cannot be assumed that the findings are wholly representative of the overall speech used by the sample in their everyday communications. It would be necessary to gather data from neighbouring age groups such as caregivers, and from a wide range of contexts for the resolution and clarification of this issue. Nevertheless, what is certain is that cultural, contextual, and linguistic influences and their interactions are accountable for the differences in the children’s early speech. Question Box: 1. What is a noun bias? Is there a noun bias in early childhood Cantonese? Why or why not? 2. What is a verb bias? Is there a verb bias in early childhood Cantonese? Why or why not? 3. Noun bias versus verb bias: Which one is more acceptable to you? Link this debate with your own observations of young children. 4. Cantonese classifiers Classifier (CL) acquisition is regarded as a crucial indicator of Chinese language learning (Tse, Li, and Leung 2007). Chinese classifiers can be categorized by the explicit features of referents, which are closely related to human cognitive categorization. This section summarizes the classifiers elicited from the corpus, and analyzes frequency of usage, the repertoire, typology, and developmental trends. 4.1. Typology of Cantonese classifiers There is some argument among linguists over the precise number of Cantonese classifiers. Although over 200 classifiers have been documented as being used by Cantonese-speaking adults in Hong Kong (Wong 1998, 2000), some researchers have proposed that there are “more than 60 classifiers in Cantonese” (Wei and Lee 2001). In addition, “the choice of classifier is often arbitrary and not predictable from the meaning or physical characteristics of the referent” (Wei and Lee 2001). Sometimes, “there may be two or more alternative classifiers for the same noun, depending on the context and on the particular attribute of the referent that the speaker wishes to emphasize,” and “these usages have to be learned individually” (Wei and Lee 2001).
38 Shek Kam Tse and Hui Li Tse, Li, and Leung (2007) identified 73 classifier types from the corpus of utterances produced by the 492 participants in the study. This figure is very close to that of the 63 proposed by Szeto (1998), who studied eight Cantonese-speaking Hong Kong children between 1.5 and 3.8 years of age. The slight difference between the two studies might be primarily due to the age difference, as the subjects in the present research were aged from 36 to 60 months. The writers believe that the total number of 73 is reasonable and acceptable, although there is no absolute consensus on the exact number of Cantonese classifiers among linguists. Allan (1977) defined classifiers as morphemes that denote salient perceived or imputed characteristics of the referents of associated nouns. He suggested four types of classifier languages: numerals, concordials, predicates, and intra-locatives. However, Aikhenvald (2000) suggested seven types of classifiers in languages across the world, which include noun classifiers, numeral classifiers, possessed classifiers, relational classifiers, verbal classifiers, locative classifiers, and deictic classifiers. According to the proposed typologies of Allan (1977) and Aikhenvald (2000), Cantonese is a numeral classifier language in which enumeration or quantification requires the use of a classifier as a bound form to designate semantic features or the quantum of whatever is being enumerated. Cantonese has a large set of classifiers, and its classifier system is intricate. Noun classifiers and verb classifiers are differentiated according to whether they classify nouns or verbs (Mak 1991; Szeto 1998; Wong 2000; Yamamoto 2005). Verb classifiers enumerate the number of times an action has taken place such as da2 saam1 kyun4 ‘hit with the fist three times’. However, the majority of Cantonese classifiers are noun classifiers (Szeto 1998). 4.1.1. Noun classifiers Lyons (1977) proposed two main types of noun classifiers, namely, (1) sortal classifiers, which belong to nouns and are classified in terms of some intrinsic features, e.g., zi1 denoting long, thin objects such as pen and gun; and (2) mensural classifiers, which denote quantities of an item, e.g., bui1 denoting a class of uncountable substances and the collective baan1 referring to a group of people.
Lexical development 39
4.1.1.1. Sortal classifiers Yamamoto (2005, p. 42) proposed “a semantic structure for Chinese classifiers”, a structure unable to encompass all Chinese classifiers, and a semantic typology of sortal classifiers per se. Despite various grammatical differences between Mandarin and Cantonese, classifiers make very similar semantic distinctions in the two varieties of Chinese. Yamamoto’s structure was employed to classify sortal classifiers in the present study. 4.1.1.2. Mensural classifiers Killingley (1983) categorized classified mensural classifiers into three subtypes: (1) collective classifiers, which form the most important and most semantically interesting class, consisting of generic classifiers (referring to a type or kind of entity, not to be confused with the “general” sortal classifier such as go3) and non-generic classifiers (denoting entities grouped in twos or units of more than two); (2) measurement classifiers, which involve the measurement of linearity (1D), surface (2D), volume (3D), weight, and other features; and (3) containment classifiers, which denote what can be contained in a box, parcel, or vessel of some kind. The typologies of Cantonese classifiers described above have been converted into a systematic typology. Both Szeto (1998) and the present study found that the core set of first classifiers produced by younger children contained all three subtypes of classifiers: numeral, verbal, and mensural. As shown in Table 2.7, all classifiers were appropriately classified into the proposed typology. The results indicate that: – Most (93.15%) of the Cantonese classifiers were noun classifiers, and few (6.85%) were verb classifiers; – Within noun classifiers, mensural classifiers exceeded sortal classifiers in terms of percentage (60.29% versus 39.71%) and total number (41 versus 27); – Inanimate classifiers (25, 92%) predominated over animate classifiers (2, 8%) within the subcategory of sortal classifiers; – Collective classifiers (14, 33.33%), containment classifiers (15, 35.71%), and measurement classifiers (12, 28.57%) had a fairly even share of the subcategory of mensural classifiers; – Concrete classifiers (22, 88%) prevailed over abstract classifiers (3, 12%) within the subcategory of inanimate classifiers;
40 Shek Kam Tse and Hui Li – There was only one classifier within the subtype of animal classifier and that of human classifier; – Even proportions of function classifiers (11, 50%) and shape classifiers (11, 50%) were found in the subcategory of concrete classifiers; – There was no predominance of 1D (4, 36.36%), 2D (4, 36%), and 3D (3, 27.27%) classifiers within the subcategory of shape classifiers. The significant predominance of noun classifiers has been reported by other Cantonese linguists (Killingley 1983). In addition to the predominance of noun classifiers, three other types of predominance were found within noun classifiers: mensural classifiers, inanimate classifiers, and concrete classifiers significantly dominated their own subcategories, respectively. The prevalence of inanimate classifiers and concrete classifiers found in adult language productions accords with the theories of other Cantonese linguists (Killingley 1983; Matthews and Yip 1994). The predominance of mensural classifiers may reflect both the specific features of Cantonese classifiers and the developmental trend of classifier acquisition. Linguistically, the relatively large percentage of mensural classifiers within the subcategory of noun classifiers may be attributed to the expanding subtype of containment classifiers. Understandably, this usage is open-ended as any container could be used as the containment classifier. At the same time, most previous studies on Cantonese classifiers have focused on sortal classifiers (e.g., Wong 2000). The relationship between the developmental trend of classifier acquisition and the finding of the prevalence of mensural classifiers are discussed in the following section.
Lexical development 41 Table 2.7. Classifiers produced by the Cantonese-speaking preschoolers I. Noun classifiers (68, 93.15%) A. Sortal classifiers (27, 39.71%) (1) Animate (2, 8.00%): Animal: zek3 ; Human: wai2 (2) Inanimate (25, 92.00%) a. Concrete (22, 88.00%) — Function (11, 50.00%): bun2, bou6, gaan1, gaa3, baa2, fo2, daam6, doe2, can1, ding2 — Shape (11, 50.00%) 1D (4, 36.36%): Flexible: tiu4; Rigid: zi1, min6 2D (4, 36.36%): fuk1, zoeng1, faai3, gaak4 3D (3, 27.27%): gau6, nap1, go3 b. Abstract (3,12.00%): joeng6, sau2, cang4 B. Mensural classifiers (41, 60.29%) (1) Collective classifiers (14, 33.33%): deoi3, fan6, bau1, pai1, paai4, zung2, baan1, hong4, tou3, zou2, di1, dip6 (2) Containment classifiers (15, 35.71%): bui1, wun2, dip6, hap6, gun3, zeon1, bou1, tung4, gang1, uk1, pun4, ci4, ci4 gang1, wu4, soeng1 (3) Measurement classifier (12, 28.57%): seoi3, hou6, man4, dou6, dim2, cyun3, cin4, hou4, jat6, jyut6, fun2, zau3, gei1, lung4 II. Verb classifiers (5, 6.85%) : ci4, leong4, kyun4, bou6, ngaan5
4.2. General classifiers General classifiers are widely used with all nouns in many numeral classifier languages (Lyons 1977). In Mandarin, ge is the most frequently used classifier. This word classifies numerous nouns; thus, it is widely regarded as the general classifier (Chao 1968; Erbaugh 1986; Li and Thompson 1981). Cantonese has the same morpheme go3, making it the most likely candidate for the general classifier. However, Mak (1991) questioned on semantic grounds whether go3 qualifies as the general classifier in Cantonese. He pointed out that go3 has its own semantic restrictions and thus cannot be used with some nouns. For instance, go3 is generally not used
42 Shek Kam Tse and Hui Li with substance-denoting nouns. Therefore, the acceptability of phrases such as jat1 go3 seoi2 ‘one (CL) water’ is marginal. Mak (1991) suggested that these phrases are more appropriate to be grouped under “mixed classifiers”, as no single classifier in Cantonese can claim to serve comprehensively as a general classifier. Although it is important to examine children’s actual use of go3 in Cantonese, the writers suggest that it is just as important to explain why a general classifier in the Cantonese acquisition process is welcome and helpful. One advantage of having a general classifier in the system is that it can predict/explain the direction of overgeneralization and why children overuse and generalize the use of go3 instead of using other specific classifiers. 4.2.1. Overgeneralizations of go3 It was found that go3 was the most frequently used classifier, with occurrence rates ranging from 86% to 96% of all the 73 classifiers used across the three age groups. The over-extended use of go3 occurs “most frequently to refer to everyday objects, which have their own specific classifiers” (Li 2001), such as human beings ni1 go3 jan4 ‘this person’ and keoi5 go3 hok6 saang1 ‘his/her student’, and round objects like jat1 go3 caang2 ‘one orange’ and loeng5 go3 bo1 ‘two balls’. Go3 was also used by the Cantonese-speaking children to replace most of the subtypes of sortal classifiers such as zek3 (animate), tiu4 (1D flexible), zi1 (1D rigid), gaa6 (function), gaan1 (function), zoeng1 (2D shape), faai3 (2D shape), po1 (3D shape), and baa2 (function). Interestingly, all the overgeneralizations of go3 occurred in the cases where a specific sortal classifier is needed to classify the noun, which indicates that, in terms of the syntactic awareness of these Cantonese preschoolers, go3 was used only to replace sortal classifiers, not all kinds of classifiers. Below are examples of children’s overgeneralization of go3. (1)
(2)
Replace zek3 (aortal, animate) a. Lo2 go3 bui1. Take CL cup. ‘Take that cup’. b. Jat1 go3 gaai1 dan2. One CL egg. ‘There is an egg’. Replace tiu4 (Sortal, 1D flexible)
Lexical development 43
a.
(3)
(4)
(5)
(6)
Li1 go3 jyu2 lei1? This (CL) fish? ‘Is this fish’? b. Li1 go3 kiu4 lai4 gaa3! This is a (CL) bridge! ‘This is a bridge’! Replace zi1 (sortal, 1D rigid) a. Zou6 go3 fo2 zin3. Make a (CL) rocket. ‘Make a rocket’. b. Bei2 go3 jik1lik6dou1 ngo3 a3. Give the (CL) Yakult to me. ‘Give the Yakult to me’. Replace gaa6 (sortal, function) a. Zing2 lan6 go3 fo2 ce1. Broke the (CL) train. ‘Broke the train’. b. Wai3 Lai4 go3 baa1 si2. Hey, give me the (CL) bus. ‘Hey, give me the bus’. Replace gaan1 (sortal, function) a. Jau5 jat1 go3 jau4 gok2 hai2 dou6. There is a (CL) post office. ‘There is a CL post office’. b. Jau5 jat1 go3 dim3 hai3 mai3 syun4 ge3 dim3. There is a (CL) shop for buying ship. ‘There is a shop for buying ship’. Replace other sortal classifiers a. Animate (horse): Ngo3 jau5 li1 go3 ma3, li1 dou6 wo3. I have this (CL) horse, here. ‘I have this horse, here’. b. 2D shape: Cai3 go3 bi4 bi1 cong4. Build a (CL) baby bed. ‘Build a baby bed’.
The children overwhelmingly favoured the classifier go3 in the early stage of language acquisition, and this finding supports Hu’s (l993a, l993b) suggestion about children’s use of classifiers. Specifically, Hu proposed
44 Shek Kam Tse and Hui Li that children acquire the syntactic structure of the numeral classifier phrase by first acquiring the general classifier as a placeholder that is later replaced by more specific classifiers. The acquisition of a general classifier is syntactically motivated, and that of specific classifiers is semantically instigated. The finding that go3 was overused by the children is pertinent to theorizing about the use of a general classifier by very young children. Chao (1968) and Erbaugh (1986) observed that the Chinese language has the general classifier ge (ge in Mandarin, go3 in Cantonese), which is applicable to any individual noun and can replace any classifier. This is in line with Lyons’ (1977) definition of a general classifier. Since then, ge has been widely accepted as the general classifier in Mandarin to classify numerous nouns (Chao 1968; Erbaugh 1986; Lyons 1977; Li and Thompson 1981). However, some researchers (e.g., Mak 1991) have suggested that go3 in Cantonese should not be treated as a general classifier, even though it can be used with a great number of nouns. The data in the present study seem to indicate that go3 in early childhood Cantonese is widely used as the general classifier. In 86% to 96% of the occasions when the children used a classifier, go3 was their first choice. They also tended to use go3 to replace only the sortal classifiers (not the mensural classifiers), including: (1) animal sortal classifiers; (2) 1D, 2D, and 3D shape sortal classifiers; and (3) function sortal classifiers. These findings are consistent with the analyses by Matthews and Yip (1994) and the findings of Fang (1985), Stokes and So (1997), Szeto (1998), and Wong (2000). Furthermore, go3 was the most frequently used classifier, even among children of four and five years of age, when they are expected to have the ability to use a greater variety of classifiers such as zoeng1, faai3, and bun2. In these circumstances, the writers suggest that go3 may be regarded as the general classifier in early childhood Cantonese. There are cognitive, linguistic, and contextual influences that may be presumed to have shaped the overuse of the general classifier by the sample. First of all, using the general classifier is a cognitively accessible and economical strategy. For example, when in the free-play sessions, the children eased the burden on their cognitive processing as they shortened response time using the widely accepted and “general” classifier. Fang (1985) noted that overuse of the general classifier might be an outcome of a cognitively accessible and economical strategy. Second, the general classifier is typically the only correct choice to refer to human other than wai2, the humanonly classifier in Cantonese. (This is a civility and is widely used in the written language.) Third, discourse factors are also important; thus, the
Lexical development 45
present study notes that the first mention of an object often features a specific classifier, but a later mention is very often simplified to the general classifier. This interpretation is given weight by the inappropriate use of other classifiers. Besides the wide use of the general classifier, zek3, tiu4 and gaan1 were also “borrowed” by the children in the present study to replace the specific classifiers. 4.2.2. Inappropriate use of other classifiers Although go3 was used very frequently, it was not the only overgeneralized classifier. Other classifiers such as zek3, tiu4, and gaan1 were also “borrowed” to replace other specific classifiers. A summary of these cases is presented below. (7)
(8)
(9)
Overgeneralizations of zek3 a. Sei3 zek3 faa1 saang1. Four (CL) peanuts. ‘Four peanuts’. (lap1 should be the classifier) b. Ngo3 gin3 dou2 zek3 se4 hai2 dou6 hang4gan2. I saw the (CL) snake walking. ‘I saw the snake walking’. (tiu4 should be the classifier) Overgeneralizations of tiu4 a. Li1 tiu4 zi3 hai6 zam1 a3. This (CL) is the needle. ‘This is the needle’. (zi1 should be the classifier) b. Li1 tiu4 mat1 je3 saam1? What this (CL) garment is. ‘What is this garment’? (gin6 should be the classifier) Overgeneralizations of gaan1 a. Ngo3 ma1mi4 hai2 gaan1 fa1 jyun2 dou6. My mum in the (CL) garden. ‘My mommy is in the garden’. (go3 should be the classifier)
As they are not general classifiers, the overgeneralization of the classifiers mentioned is regarded as “inappropriate” usage. Other instances of the inappropriate use of a classifier such as using an unnecessary classifier, ignoring the use of required classifier and so on, were not found in the present study. The finding of no omission of classifiers seems to indicate that by the age of three years, Cantonese-speaking children understand that
46 Shek Kam Tse and Hui Li a classifier is obligatory when enumerating or quantifying nouns. However, one case of using redundant classifiers was found, as evident in the following example: (10)
Utterance: Jau5 go3 jau5 go3 gaan1 maai5 tong4 gwo2 ge3 dim3 hai2 dou6. There is there is a (CL) shop selling candies over there There is a shop selling candies over there. Corrected: Jau5 gaan1 maai5 tong4 gwo2 ge3 dim3 hai2 dou6. There is a (CL) shop selling candies over there. ‘There is a shop selling candies over there.’
This example with two classifiers could be a case of self-repair or correction: the child begins with the non-target jau5 go3 (there is a CL) before repeating this sequence and eventually arrives at the target classifier, gaan1. As such, the apparent double classifier may be regarded as a performance error rather than a systematic one. The children over-generalized the three classifiers for several reasons. First of all, overusing of zek3 might be in line with Mak’s (1991) suggestion that children with higher language proficiency or who are in a higher age group tend to over-generalize specific classifiers that are commonly used in daily life. In the present study, zek3 was one of the top classifiers, and it has the second highest frequency of occurrence. Second, the overuse of tiu4 as explained by Tse, Li, and Leung (2007) in their suggestions about three types of overgeneralization in classifier acquisition in young children, is a means to: (a) substitute a verb for a classifier such as ត (jat1 piu1 wan4 ‘a piece of cloud’) where ត (piu1 ‘float’) is a verb; (b) substitute an adjective for a classifier such as ᅈН (jat1 mun3 seoi2 ‘a full [cup] of water’) where ᅈ ( mun3 ‘full of’) is an adjective; and (c) demonstrate misuse of a classifier, for example, చОً (jat1 tiu4 fo2 ce1 ‘a train’), where ӈ (lit6) should be the correct classifier instead of చ (tiu4). However, only the last type of overgeneralization was found in the children’s utterances. In many situations, the children tended to repeat the same classifier that had occurred in their previous phrases such as գԖచಭǴךԖ చً (nei3 jau5 yat1 tiu4 syun4, ngo3 jau5 yat1 tiu4 ce1 ‘you have a ship, I have a car’). In this case, they over-generalized the classifier tiu4, as they were aware that a classifier is syntactically necessary; however, they were unable or unwilling to use a more semantically appropriate one.
Lexical development 47
The overuse of gaan1 is of special interest, as there have been no similar reports in other studies. In the case of ך༰ࠓ䛦໔༜ࡋ (ngo3 ma1 mi4 hai2 gaan1 fa1 jyun2 dou6 ‘My mummy is in the garden’), the specific classifier for “garden” is ০ (zo6), which is also a function sortal classifier like gaan1. However, gaan1 is more frequently used in children’s daily communication than zo6 (around age three), which is not yet in the repertoire of Cantonese-speaking preschoolers. Possibly, this case reflects the fact that the child has not yet reached the adult level of classifier acquisition. 4.3. Semantic development and concept mastery in early childhood Although classifiers are rare in English, they are a commonplace feature of the Chinese language. Cantonese, as with other variations of Chinese, calls for a classifier after the number when referring to amounts (Matthews and Yip 1994). All nouns in Cantonese may have such classifiers associated with them that are obligatory both syntactically and semantically, for example yat1 go3 yan4 ‘one person’, leong3 tiu4 jyu2 ‘two fish’, saam1 zek3 gau2 ‘three dogs’, and sei3 gaa3 ce1 ‘four cars’. The complex syntax and semantics of classifiers pose major problems for Cantonese learners because it is difficult for them to decide which classifier to use with which noun (semantically appropriate), and when and where to use them (syntactically correct). Hu (l993a, 1993b) asserted that children acquire the syntactic structure of the numeral classifier phrase by first acquiring the general classifier as a placeholder that is later replaced by more specific classifiers. More specifically, the acquisition of a general classifier is syntactically motivated and that of specific classifiers is semantically instigated. In other words, the developmental trend is that syntactical acquisition precedes semantic mastering. The finding of the prevalence of mensural classifiers in the 492 Cantonese-speaking children may also reflect a developmental trend of classifier acquisition. It may be that developmentally, this phenomenon is associated with the maturing language proficiency of young children. The types of mensural classifiers were found significantly more than that of sortal classifiers. In contrast, nearly all the top 10 classifiers were sortal classifiers (except for one mensural classifier). This may be attributable to the developed language proficiency and can be a reflection of advancing cognitive development. Sortal classifiers qualify the noun referent according to intrinsic characteristics, whereas mensural classifiers quantify the noun refe-
48 Shek Kam Tse and Hui Li rent by a humanly determined unit (Szeto 1998). Classification by a sortal classifier is intrinsic, whereas a mensural classifier imposes an extrinsic measure on the noun referent. Hence, using sortal classifiers is more applicable than using mensural classifiers because different sortal classifiers suggest different intrinsic attributes of the noun referent. Similarly, Szeto (1998) proposed that the impact of a precise quantity on children’s classifier usage is less important than the impact of the intrinsic qualities of the noun referent. Therefore, the younger children in her study used more sortal classifiers, whereas the young children in the present study used sortal classifiers more frequently. '
Question Box: 1. What is a classifier? What is the function of classifiers in Cantonese? 2. How many types of classifiers were identified in the early childhood Cantonese? What are they? 3. Are there any over-generalized classifiers in early childhood Cantonese? If so, what are they? Why are these classifiers overused? What are the implications for Cantonese learning and teaching?
Lexical development 49
5. Conclusion and implications The typology and repertoire of Cantonese classifiers found in the present study set a cornerstone for further theory building. The discovery that young Cantonese-speaking children tend to over-generalize several commonly used classifiers should be noted by early childhood educators. Such usage should be tolerated, and children should be provided with structured learning to utilize and extend their linguistic repertoire. Specifically, teachers should understand that such an overgeneralization is a normal process or stage as children acquire the complex system of Cantonese classifiers. Therefore, teachers should ask children to use classifiers more frequently in their in-class conversations. Teachers should also provide well-planned learning for children, in which the different Cantonese classifiers are systematically presented and explained to children. Furthermore, it is particularly important for teachers to understand the use of the various Cantonese classifiers and to avoid the use of inappropriate classifiers in their interactions with children. The findings that verbs dominate Cantonese-speaking children’s early utterances may have some contribution to theory building and may have practical implications for early language education. Specifically, the finding that Cantonese speakers display a predominance of verbs in their language in early childhood may indicate a clear direction for language teachers in kindergartens. It is suggested that Hong Kong preschool teachers replace the extensive copying, dictation, and rote-memory exercises that predominate the kindergarten classroom (Ho and Bryant 1997; Opper 1992) with the active teaching approach, involving more activity, movement, role play, and other related activities in the teaching of Cantonese. By doing so, children will be exposed to a variety of verbs in their learning, which will help facilitate their language development. In light of the finding that the participating children did not display a predominance of nouns in their utterances, teachers may consider providing the opportunity for children to acquire more nouns.
Chapter 3 Syntactic development in the early years
Syntax refers to the grammatical arrangement of words in sentences. A relatively small number of in-depth studies have investigated children’s acquisition of Cantonese and how syntactic skills or the ability to structure sentences properly develops in Cantonese-speaking children. Kwong (1990) and Lee, Wong, Leung, Man, Cheung, Szeto, and Wong (1996) have conducted small-scale studies on the developmental path of syntactic competence in Chinese-speaking children. This chapter discusses the syntactic development in a larger sample of children than those examined in previous studies, focusing in particular on (a) the average number of utterances recorded per child at different ages, (b) the mean length of utterances (MLU), (c) the range and frequency of use of sentence patterns in the children’s Cantonese, (d) the types of declarative sentences that were regularly used, (e) the syntactic complexity of simple declaratives, and (f) the verb patterns of declarative sentences. In addition, the developmental trends and their implications will be discussed in the last section. 1. Indicators of syntactic development Syntactic development is usually gauged in terms of changes in MLU, sentence type and structure, syntactic complexity, and verb pattern. Psycholinguistic research on syntactic development in Western languages has tended to focus on the acquisition of syntactic rules and comprehension of specific subsystems such as coordination, passives, and various word-order combinations. MLU is found to be a reliable indicator (Bowerman 1979; MacWhinney 1982) of syntactic development. Therefore, the present study employed MLU and syntactic complexity as the key measures to understand syntactic development in Cantonese-speaking preschoolers.
Syntactic development 51
1.1. Mean length of utterances Brown (1973) was among the first researchers to use MLU to track down syntactic development in the English of young American children, using the total number of morphemes (units of meaning) rather than words to indicate the length of utterances produced in natural contexts. Given the nature of written and spoken Chinese and the fact that there is no graphemic transformation in Chinese, investigators of the Chinese language have been compelled to use either the number of characters in transcriptions or the number of words (usually composed of two or more characters) to indicate sentence length and MLU. The application of different foci and use of methods of calculating MLU in Chinese studies have led to a degree of inconsistency in estimates of MLU for children at various ages. For example, Chen and Ryback (1974) reported that the MLU at 32 months was 7.7 morphemes; Wu and Xu (1979) reported 15 characters at age three; whereas Kwong (1990) proposed an average of 3.4 words for three-year-old Cantonese-speaking children. In the present study, we used the number of characters of each utterance in transcriptions to indicate the length of utterance. Regardless of the method used for calculating MLU, the length of sentences uttered has been found to increase with the age of the child during preschool years (Chen 1995; Jin 1994). Although MLU provides a useful reference point and reflection of children’s syntactic development, there are problems associated with using MLU as the fundamental indicator of syntactic development in children. MLU itself does not provide a comprehensive measure of children’s mastery of syntax, as sentences having the same MLU may possess quite different levels of syntactic complexity. There is also a lack of a commonly recognized definition of what constitutes a “word” in some non-morphemic languages, making it even more difficult for researchers to arrive at a universal coding and assessment scheme for MLU. Furthermore, MLU is considered more reliable for younger ages. Given these issues associated with the use of MLU, syntactic complexity was utilized in the present study as another key index of growth in the development of syntax of the Cantonese-speaking children investigated.
52 Shek Kam Tse and Hui Li 1.2. Syntactic complexity Syntactic complexity, as the key indicator of syntactic development, has been widely drawn upon in studies of the Chinese language. Both the Cantonese-based study by Kwong (1990) and the Mandarin-based study of syntactic complexity by Zhu (1979) found that sentences become more refined as children grow older, with increases in the complexity of modifiers such as subject predicate as object or subject. Kwong (1990) reported a significant spurt in syntactic complexity in the sentences of four-year-olds, with no such marked increase appearing in the sentences of children between the ages of four and five. Kwong found that between the ages of three and four, the proportion of declarative sentences with simple modifiers increased significantly, achieving 60% of all utterances in the speech of children by the age of five. This figure is much lower than the 85.2% figure for fiveyear-olds reported in Zhu’s study of Mandarin-speaking children. Such a difference in proportion may be a consequence of the differences in sampling approach as well as the differences between Cantonese and Mandarin. The relatively small sample size in Kwong’s study (60 children aged from three to five years) also limits the generalizability of her findings; thus, further research with a larger and more representative sample in Hong Kong was needed to corroborate her conclusions. The present study used various sources for deciding the units of analysis (Kwong 1990; Scarborough 1985; Zhang 1985; Zhu 1986). Growth in syntactic development was measured in terms of the frequency and percentage of utterance length, sentence type and structure, syntactic complexity, and verb pattern in declarative sentences. The framework used in the analysis is shown in Table 3.1.
Syntactic development 53 Table 3.1. Framework of utterance analysis Variable
Attribute
Length of utterances
— Number of utterances
Sentence types
— MLU — Declarative sentence — Interrogative sentence — Exclamation sentence — Negation sentence — Imperative sentence — Incomplete sentence — English and other words
Structures of declaratives
— Single-word sentence — Simple declarative — Subjectless sentence — Subject-predicate sentence — Compound sentence
Syntactic complexity of simple declaratives
— Without modifier — With simple modifier — With complex modifier — With subject-predicate (SP) as subject or object — With intransitive verb — With transitive verb — With co-verbs or verbs in serial expression modifier — Link verb — Without verb
Verb patterns in declaratives
Question Box: 1. What is MLU? Are there any differences between Cantonese and English regarding the definition of MLU? How should one measure MLU in Cantonese? 2. How many types of sentences can you identify in early childhood Cantonese? What are they? Can you cite one example for each type? 3. What is syntactic complexity? How may it best be measured in early childhood Cantonese?
54 Shek Kam Tse and Hui Li 2. Syntactic development in the early years 2.1. Mean length of utterance Considering the limited resources and manpower, we randomly selected only 180 children (60 children from each age group) from the corpus of 492 children. The sentences of the 180 children composed the database for the analysis, allowing time for the precise calculation of MLU and the number of Chinese characters in the children’s utterances. The mean number of utterances calculated was found to be 48.6, with a standard deviation of 14.29 utterances. The mean of the utterances of three-year-olds (M = 47.88, SD = 14.39) did not seem to differ much from that of the four-yearolds (M = 47.50, SD = 13.12). However, five-year-olds produced more utterances than the younger children (M = 50.57, SD = 15.31). The mean and standard deviation of the children’s utterances are shown in Figure 3.1.
Figure 3.1. Mean of utterances across the three age groups
Although the differences in the mean of utterances did not appear to be significant, further analyses show a steady and continuous growth in the MLUs of the three age groups. The MLU of the three-year-olds was 5.72 (SD = 1.04); however, this increased by 4% to 5.96 (SD = 1.11) in fouryear-olds, and increased by another 6% to 6.4 (SD = 1.36) in five-year-olds. In fact, the growth in the MLU between four- and five-year-olds was found
Syntactic development 55
to be statistically significant, p