中文漢字對日語漢字解讀的影響

全文

(1)國立臺灣師範大學英語學系碩士論文 Master’s Thesis Department of English National Taiwan Normal University. 中文漢字對日語漢字解讀的影響 The Influence of L1 Chinese Characters on the Interpretation of Japanese Kanji. 指導教授：陳純音博士 Advisor: Dr. Chun-yin Doris Chen 研究生：羅佳琪 Student: Jia-Qi Jan Luo. 中華民國一百零四年六月 June 2015.

(2) 摘要本研究旨在探討以中文為母語的日語學習者對日文漢字形義之第二語言解讀情形，主要研究的議題包括第一語言遷移現象、漢字形上的對應、漢字意義上的對應、難易次序、題型效應以及日語能力對漢字解讀的影響。本研究採用兩個實驗題型：漢字圖示對應題與漢字文句對應題，試題皆以選擇題的形式呈現。受試者為七十二位以中文為母語的日語系學生以及二十四位日語母語人士，日語系學生依其日語標準化測驗，分為初、中、高三組。首先，研究發現，受試者在解讀日語漢字時受到第一語言遷移的影響。在形上，受試者的表現在同形與半同形上比異形表現要來得好，此結果顯示中文漢字與日文的形式對應能幫助日語學習者解讀。針對字義，同義又比異義的表現好，顯示出字義的對應也能幫助日語學者學習。就類別而言，受試者在同形同義與半同形同義上比同形異義與半同形異義還要好，此結果顯示字義的對應比字形的對應還要更重要。在題型效應方面，受試者在漢字文句對應的表現比漢字圖示對應的表現要好，說明對於整體受試者，在判斷漢字時，文句中的線索比圖示中的線索還來得有效。最後在語言程度方面，受試者之對於日語漢字的解讀會隨著日語程度的提升而改善。. 關鍵詞：第一語言遷移、漢字、字形、字義、難易次序、題型效應. i.

(3) ABSTRACT The present thesis aims to investigate the interpretation of kanji characters by Chinese learners of Japanese in Taiwan. The discussed issues of this study include L1 transfer, form-driven effects, meaning-driven effects, the difficulty order of kanji characters, task effects, and proficiency effects. A kanji-picture mapping task and a kanji-text mapping task were designed, both of which were presented in multiple choice format. Seventy-two Japanese majors were recruited and further divided into three experimental groups, according to their performances on a standardized proficiency test in Japanese, basic, intermediate and advanced. In addition, twenty-four Japanese native speakers participated in the present study as our baseline group. The overall results indicate that L1 positive transfer was found influential in the L2 acquisition of Japanese kanji characters. With regard to the forms of kanji characters, the participants performed better on same forms and semi-same forms than on different forms. The form correspondence between Chinese characters and kanji characters was indeed helpful. With regard to the meanings of kanji characters, the participants’ performance on same meanings was better than different meanings, indicating that the meaning correspondence could help L2 learners acquire kanji characters as well. Concerning the difficulty order of the kanji characters, it was found that SF-SM (same form and same meaning) and SsF-SM (semi-same form and same meaning) were easier than SF-DM (same form and different meaning) and SsF-DM (semi-same form and different meaning). The results showed that the meanings of kanji characters were more dominant than the forms. In response to the task effects, the participants performed better on kanji-text mapping task than on kanji-picture mapping task. The textual cues were more efficient than pictorial cues for our L2 learners. Finally, the interpretation of Japanese kanji characters was found to be influenced by L2 learners’ proficiency. The higher proficient a Japanese learner was, the ii.

(4) better he interpreted kanji characters.. Keywords: L1 transfer, kanji characters, form, meaning, difficulty order, task effects. iii.

(5) ACKNOWLEDGMENTS The completion of this thesis doubtlessly needs lots of help from many people. Without these people’s assistance, I could not imagine that I could come to this stage. First of all, I would like to express my grateful gratitude to my thesis advisor, Dr. Chun-yin Doris Chen who led me into the field of second language acquisition. Dr. Chen was thoughtful and inspired. Since I was in her class, her inspired expressions impressed me, and hence I was motivated to work on L2 acquisition. In the process of my thesis writing, her helpful guidance and suggestions helped me to develop and organized the thesis. When I was depressed on my thesis, she always encouraged me. Her patience and optimism made me keep working on the thesis with confidence. Without her, this thesis could not have been possible. I am also grateful to my considerate committee members, Dr. Jyun-gwang Fred Chen and Dr. Rui-ling Sharon Fahn. Their insightful remarks and constructive advice help me improve the quality of this thesis. Special thanks go to Dr. Shu-juan Chen, Dr. Sun-bun Zhao, Office of International Affairs of NTNU, Shu-feng Li (my Japanese teacher at NUTN). Without their kindly help, I could not have recruited enough subjects and collect data efficiently. Sincere appreciation is also given to the teachers who taught me during the graduate studies at the Department of English of NTNU: Dr. Gerardo Fernández-Salgueiro, Dr. Hui-shan Nissa Lin, Dr. Jing-lan Joy Wu, Dr. Miao-hsia Tammy Chang, Dr. Wen-da Zeng, Dr. Xiao-hong Iris Wu, Dr. Xiao-hui Zhan, Dr. Zhen-yi Li by alphabetic order. With their teaching, I could explore the diverse aspects of linguistics. I benefit considerably from their instruction. Furthermore, my gratitude is extended to all my classmates of the of the MA Program in Linguistics in NTNU: Bill Chang, Doris Chen, Ivy Zeng, Jennifer Hsu, John Lu, Karen iv.

(6) Kan, Ken Lin, Kevin Song, Lilian Chen, Lucy Chiang, Vicky Lin by alphabetic order. I gained much benefit from their knowledge, experiences, and friendship. Sincere appreciation is also given to my warmhearted great friends. I thank Meng-qi Zhou, Christine Li, Abbie Zhang for listening to my stories of daily life. I also thank to Jocelyn Chen and Cosin Jiang for giving me assistance and sharing happiness with me as school sisters. Moreover, I want to extend my thanks to staffs at Department of English of NTNU: Hao-ku Bartleby Hsu, Hui-chiao Sun, Mu-han Wang, Ping-jung Ho, Shu-hui Wang, Tzu-yi Jiang by alphabetic order. As staff members, they care us and also gave me great assistance. I am fortunate to meet them. Last and most of all, I am grateful to my family for their unconditional support, understanding and encouragement. They give me the greatest love, and they always believe in me. I would like to dedicate this thesis to them.. v.

(7) TABLE OF CONTENTS CHINESE ABSTRACT .......................................................................................................... i ENGLISH ABSTRACT ......................................................................................................... ii ACKNOWLEDGEMENTS .................................................................................................. iv TABLE OF CONTENTS ...................................................................................................... vi LIST OF TABLES .............................................................................................................. viii LIST OF FIGURES ............................................................................................................... x CHAPTER ONE INTRODUCTION ................................................................................................ 1 1.1 Motivation ................................................................................................................... 1 1.2 Theoretical Background .............................................................................................. 4 1.2.1 Form Formation ................................................................................................... 4 1.2.2 Meaning Formation ............................................................................................. 5 1.2.3 Contrastive Analysis: Difficulty Order ............................................................... 6 1.3 Research Questions ..................................................................................................... 8 1.4 Significance of the Study .......................................................................................... 10 1.5 Organization of the Thesis ........................................................................................ 10 CHAPTER TWO LITERATURE REVIEW.................................................................................... 11 2.1 Previous Theoretical Studies of Japanese Kana and Kanji ....................................... 11 2.1.1 Leong & Tamaoka (1998) ................................................................................. 11 2.1.2 Kess & Miyamoto (1999) .................................................................................. 14 2.1.3 Nagano & Shimada (2014) ................................................................................ 18 2.1.4 Summary............................................................................................................ 21 2.2 Previous Empirical Studies of Kanji ......................................................................... 22 2.2.1 Chikamatsu (1996) ............................................................................................ 22 2.2.2 Mori (2003) ....................................................................................................... 25 2.2.3 Haththotuwa Gamage (2003) ............................................................................ 27 2.2.4 Toyoda (2009) ................................................................................................... 29 2.2.5 Summary............................................................................................................ 31 2.3 Summary of Chapter Two ......................................................................................... 34 CHAPTER THREE RESEARCH DESIGN .................................................................................... 35 3.1 Participants ................................................................................................................ 35 3.2 Methods and Materials .............................................................................................. 36 3.2.1 A New Classification of Kanji Characters ........................................................ 37 3.2.2 Two Tasks ......................................................................................................... 42 3.2.3 The Kanji-picture Mapping Task ...................................................................... 44 vi.

(8) 3.2.4 The Kanji-text Mapping Task ........................................................................... 45 3.3 Procedures ................................................................................................................. 45 3.3.1 The Pilot Study .................................................................................................. 46 3.3.2 The Formal Study .............................................................................................. 47 3.3.3 Scoring and Statistical Analysis ........................................................................ 48 3.4 Summary of Chapter Three ....................................................................................... 48 CHAPTER FOUR RESULTS AND DISCUSSION .......................................................................... 50 4.1 Form-driven Effects .................................................................................................. 50 4.2 Meaning-driven Effects ............................................................................................. 58 4.3 Type Effects .............................................................................................................. 64 4.4 Task Effects ............................................................................................................... 72 4.5 Proficiency Effects .................................................................................................... 79 4.6 Summary of Chapter Four ......................................................................................... 81 CHAPTER FIVE CONCLUSION ................................................................................................ 82 5.1 Summary of the Major Findings ............................................................................... 82 5.2 Pedagogical Implications .......................................................................................... 83 5.3 Limitations of the Present Study and Suggestions for Future Research ................... 84. Bibliography ........................................................................................................................ 85 Appendix A Appendix B Appendix C Appendix D. Items Used in Task 1: Kanji-picture Mapping Task .................................. 91 Items Used in Task 2: Kanji-text Mapping Task........................................ 99 Results of the Pilot Study ......................................................................... 102 Consent Form ........................................................................................... 104. vii.

(9) LIST OF TABLES Table 1-1 Difficulty Patterns and Levels of Difficulty .......................................................... 7 Table 1-2 Hierarchy of Difficulty ........................................................................................ 8 Table 2-1 A Summary of the Major Findings and Limitations ........................................... 32 Table 3-1 A Summary of the Participants............................................................................ 36 Table 3-2 A Summary of the New Classification of Kanji Characters ................................ 41 Table 3-3 A Summary of the Task Design .......................................................................... 43 Table 3-4 A Test Example in Pictorial Format .................................................................... 44 Table 3-5 A Test Example in Textual Format ..................................................................... 45 Table 4-1 Participants’ Overall Performances on the Three Types of Forms of Japanese Kanji Characters ................................................................................... 50 Table 4-2 Participants’ Overall Performances on the Two Types of Forms (combining same and semi-same forms) of Japanese Kanji Characters ............. 51 Table 4-3 The p-values for the within-group Differences among Three Types of Forms of Japanese Kanji Characters ................................................................... 51 Table 4-4 The Results of the Scheffé Post Hoc Analysis (Comparison between Three Types of Forms) ................................................................................................... 52 Table 4-5 Hierarchy of Difficulty ........................................................................................ 53 Table 4-6 Revised Hierarchy of Difficulty and the Examples from the Tasks .................... 54 Table 4-7 The p-values for the Three Types of Forms of Japanese Kanji Characters......... 57 Table 4-8 Participants’ Overall Performances on the Same and Different Meanings of Japanese Kanji Characters ................................................................................... 59 Table 4-9 The p-values for the within-group Differences between Same and Different Meanings of Japanese Kanji Characters .............................................................. 59 Table 4-10 The p-values for the Same and Different Meanings of Japanese Kanji Characters.......................................................................................................... 61 Table 4-11 The Resutls of Scheffé Post Hoc Analysis (Comparison within Two Types of Meanings) ..................................................................................................... 62 Table 4-12 Participants’ Overall Performances on the Two Types of Japanese Kanji Characters.......................................................................................................... 63 Table 4-13 The p-values for the within-group Differences among Two Types of Japanese Kanji Characters ................................................................................ 64 Table 4-14 Corresponding Forms and Meanings in the Revised Hierarchy of Difficulty ........................................................................................................... 64 Table 4-15 Participants’ Overall Performances on the Four Form-Meaning-Integrated Types of Japanese Kanji Characters....................... 65. viii.

(10) Table 4-16 The p-values for the within-group Differences among Four Form-Meaning-Integrated Types of Japanese Kanji Characters....................... 65 Table 4-17 The Results of Scheffé Post Hoc Analysis (Comparison between Four Form-Meaning- Integrated Types) .................................................................... 66 Table 4-18 The p-values for the Four Form-Meaning-integrated Types of Japanese Kanji Characters ................................................................................................ 69 Table 4-19 The Results of Scheffé Post Hoc Analysis (Comparison within Four Form-Meaning-Integrated Types) ..................................................................... 70 Table 4-20 Examples of the Test Items in the PM Task and the TM Task ......................... 72 Table 4-21 Participants’ Overall Performances on the PM Task and the TM Task ............ 73 Table 4-22 The p-values for the within-group Differences on the PM Task and the TM Task ............................................................................................................ 73 Table 4-23 The p-values for the PM Task and the TM Task ............................................... 74 Table 4-24 The Results of Scheffé Post Hoc Analysis (Comparison within Two Tasks) .... 75. ix.

(11) LIST OF FIGURES Figure 3-1 The Procedure of the Present Study ................................................................... 48 Figure 4-1 Overall Performances of Each Group on the Three Types of Forms of Japanese Kanji Characters ................................................................................ 57 Figure 4-2 Overall Performances of Each Group on the Same and the Different Meanings of Japanese Kanji Characters ........................................................... 61 Figure 4-3 Overall Performances of Each Group on the Four Form-MeaningIntegrated Types of Japanese Kanji Characters ................................................ 69 Figure 4-4 Overall Performances of Each Group on the PM Task and the TM Task ......... 73 Figure 4-5 Participants’ Performances on the Three Types of Forms of Japanese Kanji Characters in the PM Task and in the TM Task ...................................... 76 Figure 4-6 Participants’ Performances on the Two Types of Meanings of Japanese Kanji Characters in the PM Task and in the TM Task ...................................... 78 Figure 4-7 The L2 Participants’ Developmental Stages of the Japanese Kanji Characters.......................................................................................................... 79. x.

(12) CHAPTER ONE INTRODUCTION In this chapter, motivation, theoretical backgrounds, research questions, significance of the present study and organization of the thesis are presented. Section 1.1 provides the motivation of the present thesis, and Section 1.2 introduces theoretical backgrounds from three studies on issues of form, meaning and learner strategies. The research questions are proposed in Section 1.3. Section 1.4 is the organization of the thesis.. 1.1 Motivation Studies on second language acquisition have expanded in quantity as researchers developed different theories of the second language acquisition, such as Contrastive Hypothesis or Identity Hypothesis (Klein, 1946), and also discussed in various languages. Many studies focused on the acquisition of English (e.g. Ionin, Zubizarreta, & Maldonado, 2008) or other European languages (Kanno, 1999). However, acquiring Japanese as a second. language has become popular since the last two decades (Iwasaki, 2004). To date, studies on the acquisition of Japanese increase and they focus on different linguistic areas, including phonology and syntax. For example, in L2 phonology, Toda (2003) investigated the development of syllable timing system of Japanese by English speakers. In L2 syntax, Hirakawa (2001) examined English-speaking learners’ acquisition of Japanese unaccusative verbs. The aspects of semantics are also included in Hirakawa’s discussion, such as telicity and change of state. In a more specific area of Japanese, kanji is one of the interesting topics for many researchers (Epp, 1969; Hirose, 1998; Klingborg, 2012). There were studies conducted on the phonological, semantic and neurolinguistics aspects of kanji (Morita & Matsuda, 2000; Nagahara, Amagase, & Hatta, 2006; Saito, Yamazaki, & Matsuda, 2002).. 1.

(13) However, most of them investigated kanji acquisition by English-speaking learners or other European-language learners. Few studies have been conducted on the acquisition of kanji by Chinese learners of Japanese. It has been generally known that Japanese is unique in that it has four writing systems, including hiragana, katakana, kanji and romazi (Kess & Miyamoto, 1999). The Japanese reading system is based on the kana (hiragana and katakana). In the Japanese phonetic system, its smallest phonetic element is an open syllable and it usually contains a vowel or one or two consonants with a vowel, such as あ /a/, す /su/, and つ /tsu/ (Ohata, 2004). As for the Japanese basic grammatical structure, it is an SOV language. Tae (2012) provided a general view of Japanese grammar that nouns usually accompany with particles. は /wa/ in せんせいは introduces the noun せんせい ‘teacher’ as a topic in the sentence. If the noun is followed by が /ga/, が functions as a subject marker in syntax. Besides, if the noun is the direct object of a verb, it would be followed by を /wo/ as in ご飯をたべます ‘to eat dinner’. Its declarative sentence usually ends with V-ます/masu/ in formal speech to make the sentence sound more polite. A Japanese interrogative sentence ends with か, which is also a question marker in its polite form. In addition to Japanese general grammar structure, kanji, as a part of the Japanese written system, is written in the form of Chinese characters. Nouns, stems of adjectives, and verbs are three types of words in Japanese that are almost written in kanji form. The reason why Japanese speakers use kanji is that it is useful to help them distinguish meanings when the pronunciations of the words are same1. In fact, Japanese kanji and Chinese characters are in common in terms of form and meaning since both languages are logograms. Languages with logographic system do not 1. In terms of history, Japanese kanji characters were borrowed from Chinese in different periods. From the perspective of form, they developed their own kanji but also preserved some Chinese-originated characters. Thus, we can find form correspondences or meaning correspondences between Japanese kanji and Chinese characters. Readings are related as well because they also preserved Chinese reading as on-reading. Moreover, according to Moro Miya’s “Hanzi Riben” (漢字日本, 2014), one of the reasons that Chinese and Japanese have these differences in form and meaning is because Japanese speakers misinterpreted the meanings of those borrowed Chinese characters. 2.

(14) have one-to-one correspondence between sounds and symbols. A combination of two or more phonemes can correspond to one or more logographs (e.g.ち /chi/ for 地 or 血). The forms of these logographs are very similar to or exactly the same as Chinese characters. Kanji characters are also meaning-related. When we form a two-kanji character, each kanji character is assigned a meaning before two characters are combined (Vee, 2006). Usually, Japanese kanji characters accompany kana words. The meanings of those kana words can be derived from the meanings of their kanji elements, as can be seen in (1). (1) a. 熱い /atsu-i/ ‘hot’ b. 綺麗 /kirei/ ‘beautiful; clean’ c. 案外 /angai/ ‘surprisingly’ In (1a) - (1c), the forms and meanings of characters 熱/ 綺麗/ 案外 in Japanese are the same forms as the forms of Chinese characters. Chinese learners may assume that these kanji characters and Chinese characters share the same meaning. In (1a), learners might think that 熱 means ‘hot’ since it is the meaning in Chinese. It is true that Japanese 熱 and Chinese 熱 are same in meaning. However, (1b) and (1c) are different cases. In (1b), Chinese learners of Japanese might think that 綺麗 in Japanese means ‘beautiful’ only but it also means ‘clean’ in Japanese, which is not a possible meaning for 綺麗 in Chinese. As for (1c), the meaning of 案外 in Japanese is very different from the meaning of Chinese words. In Chinese we may translate 案外 as ‘(something/event) out of a case’. However, Japanese 案外 means ‘surprisingly’. Chinese learners of Japanese are highly likely to struggle in this kind of kanji characters because the forms are same as Chinese characters but their meanings are different. The adoption of first language knowledge in L2 acquisition is L1 transfer (Ellis, 1997). Since L1 transfer may occur when Chinese students learn Japanese kanji, in this thesis I would like to see how L1 transfer would influence Chinese learners of Japanese. Furthermore, proficiency is a concern in this study as well. In some studies, it has been indicated that proficiency might affect second language acquisition (e.g. Fernández Dobao, 3.

(15) 2002; Toyoda, 2009). According to Tremblay (2011), a standardized assessment of proficiency is important and it should not be optional in second language research. Therefore, studying kanji acquisition, proficiency can be an influential factor. For example, learners at the basic level may struggle with the meaning of kanji characters which have the same form in Chinese but are different in meaning. While they become intermediate or advanced learners, their interpretation of kanji may be different. Therefore, the current study aims to examine the L1 transfer effect and proficiency effect on Chinese learners’ acquisition of kanji. As the forms and meanings of the Japanese kanji characters and the Chinese characters share some similarities, Chinese learners may transfer their L1 knowledge into L2. The correspondences with form and meaning are considered two important factors in affecting their kanji acquisition. How these two factors will influence Chinese learners of Japanese is the focus in the present study.. 1.2 Theoretical Background This section introduces some theoretical backgrounds on Japanese kanji. As mentioned above, form and meaning are two important issues that will be addressed in the present study. Therefore, Sections 1.2.1 and 1.2.2 will provide some concepts with meaning and form of kanji. The background information about learner strategies will be presented in Section 1.2.3.. 1.2.1 Form Formation In Japanese, there are four major types of forms (He, 2004). The first type refers to Japanese characters (国字 Guozi). The definition of Japanese characters is controversial, but the Japanese characters refer mainly to the characters which are Japanese-originated, such as 込. These characters are not borrowed from Chinese and they can only be read in. 4.

(16) kun-reading2. The second type is Japanese simplified characters. Some of these Japanese simplified characters are same as modern simplified Chinese but some are not. Some of these characters are derived from ancient Chinese. While these ancient simplified characters are commonly used in modern Japanese, they are not used in Modern Chinese. For instance, 拝 is listed in 1,945 jooyoo kanji (The Agency for Culture Affair of Japan, 2010) but in modern Chinese 拜 is used. The third type is popular Japanese characters (日本俗字 Riben suzi). Compared to the standard characters in each ancient period, these Japanese characters are the most popular words used in lower society. For example, 包 can be used as a popular form of 庖. The last type is Hanzici (or Sinoxenic), such as 書. The Hanzici refers to words only written in kanji, no matter they are kun-reading or on-reading. This type of kanji includes Chinese-originated characters and Korean-originated characters. However, according to Yamashita and Maru (2000), there is another type of kanji, katakana composites. Katakana composites are formed by katakana characters, such as タ and ト formed kanji 外.. 1.2.2 Meaning Formation Japanese kanji characters are meaning-related as are Chinese characters (Halpern, 2001). A kanji character may have different senses or meanings. The meanings of a kanji can be operated at four levels, as in (3). (3) a. on independent word (e.g. 明 ‘discernment’ in 先見の明 'foresight') b. on word element (e.g. 明 'clear, obvious' in 明確な 'clear, distinct') c. an independent kun word (e.g. 明るい 'bright, light') d. a kun word element (e.g. 明け 'end' in 忌明け 'end of mourning'). 2. Kanji characters are derived from Chinese characters. This exchange of Chinese characters has gone through several periods of time and the pronunciation of the Chinese characters might be preserved or might change. Hence, Japanese kanji characters can be read in two readings, on-reading and kun-reading. The Chinese readings (or Sino-Japanese readings) are called on-reading and Japanese readings (or native readings) are called kun-reading. For consistency, this thesis uses on-readings and kun-readings. 5.

(17) Sometimes, a kanji character can have different meanings at one or more than one level. These levels may then interact in a complex way. Moreover, a kanji character may function as a morpheme. It may have a bound form and a free form. Hence, when kanji characters are combined, numerous meanings of kanji may be generated. For example, the on word element 山 /san/ and the kun word element 山 /yama/ have the same meaning as ‘mountain’. However, the on word element 山 /san/ also has a sense of ‘Buddhist temple’ in 本山 /honzan/. Therefore, this complex interaction in kanji combination may cause learning difficulties for L2 learners of Japanese. As mentioned by Vee (2006), kanji characters are meaning-related in nature. However, the question is whether kanji characters only represent meanings. According to Matsunaga (1996), kanji characters commonly refer to pictographs, ideographs, logographs, or morphographs. Those who use these terms may claim that kanji characters represent meanings independent of sounds. However, Matsunaga reexamined kanji and found that kanji characters are not only meaning-related but also sound-related. He, thus, considered kanji “morphophonemic”. The nature of kanji is still controversial.. 1.2.3 Contrastive Analysis: Difficulty Order The theory of contrastive analysis hypothesis assumes that the degree of difference between the target language and the source language will affect the L2 learners’ acquisition. The more different the two languages are, the more difficult L2 learners perceive. As advanced by Lado (1957), he describes the relation between first language and second language with the notions of similar and different on forms and meanings. Difficulty patterns are proposed in response to how to compare two vocabulary systems, including Cognates, Deceptive Cognates, Different Forms, Strange Meanings, New Form Types, Different Connotation and Geographical Restrictions, as shown in Table 1-1.. 6.

(18) Table 1-1 Difficulty Patterns and Levels of Difficulty Levels. Difficulty Patterns. Easy. Cognates. Normal. Different Forms. Difficult. Deceptive Cognates. Difficult. Strange Meanings. Difficult. New Form Types,. Difficult. Different Connotation. Difficult. Geographical Restrictions. The Cognates refers to the words that are similar in forms and meanings in the L1 and L2. The words in L2 may be borrowed from the L1. The difficulty level of this pattern is identified as easy. The words that are different in forms but similar in meanings are identified as the Different Forms and the difficulty level is classified as normal. The Deceptive Cognates are the words that are similar in forms but different in meanings. This kind of words in the L2 may also be borrowed from the L1 but meanings changed in L2. The difficulty level of this pattern is identified as difficult. The Strange Meanings refers to the words that are different in forms and meanings. The different meanings in this pattern are “strange” (Lado, 1957, p.85) for the L1 speakers in that a given concept is interpreted differently in the two languages. Thus, the difficulty level of this pattern is difficult. The New Form Types refers to the words that are different in the morphological construction (Lado, 1957, p.86) and the difficulty level of this pattern is recognized as difficult. If the words that are very different in connotations in the two languages, these words will be identified as Different Connotation. The difficulty level of this pattern is also difficult. The last Geographical Restrictions refers to the words that are only used in certain geographical area and the difficulty level is difficult as well. Stockwell, Bowen and Martin (1965) also agree that the differences between the L1 and L2 will be challenging for L2 learners. Thus, he takes grammatical differences between. 7.

(19) English and Spanish as examples and proposes his own five-type hierarchy of difficulty (simplified version, adapted from Larsen-Freeman & Long, 1991, p.54).. Table 1-2 Hierarchy of Difficulty Degree. Type of Difficulty. Easy. Correspondence Coalesced Absent New. Hard. Split. The Correspondence was the easiest, whereas the Split was the most difficult. The Correspondence type refers to a function or a grammatical structure which exists in both L1 and L2. The Coalesced means that different grammatical structures with multiple meanings in the L1 are realized as a single grammatical structure in the L2. The Absent occurs when a form in the L1 does not have the correspondence in the L2. The fourth type, the New, means that a function or a structure only exists in the L2 but no correspondence in the L1. Finally, the Split means that a grammatical structure in the L1 is split into two or more grammatical structures with multiple meanings in the L2.. 1.3 Research Questions Based on the previous literature and the theoretical background displayed in the previous subsections, the present study aims to investigate the following five research questions. 1. Does the form correspondence between Japanese kanji characters and Chinese characters help Chinese learners of Japanese to gain a better understanding of kanji meaning? 2. Does the meaning correspondence between Japanese kanji characters and Chinese characters help Chinese learners of Japanese to perform better? 8.

(20) 3. How does the form interact with the meaning and what is the difficulty order of the kanji types? 4. Do different task formats elicit different experimental results from Chinese learners of Japanese when they interpret Japanese kanji characters? 5. Do advanced learners perform better on Japanese kanji recognition? 6. Is there L1 transfer (positive or negative transfer) in the Chinese second language learners’ acquisition of Japanese kanji? The first and second research questions discuss whether kanji characters of different form and with different meaning correspondences will affect L2 learners’ recognition of kanji. As kanji in the aspects of form and meaning is very similar to Chinese and is meaning-based (Vee, 2006), it is expected that with the aid of form and meaning, learners can reach better performance when inferring meanings. The third research question examines the type effect and how different types of kanji characters will affect Chinese learners of Japanese on interpreting kanji characters. If the difficulty order can be found, the study of Lado (1957) and Stockwell et al. (1965) will be supported. The fourth research question addresses the differences between textual cues and pictorial cues and how participants interpret kanji from two different tasks. The fifth question examines the proficiency effect in the learning of kanji. In the present study, it is expected that learners at different proficiency levels will perform differently on tasks. If such performance differences are found, it will support the proficiency effect mentioned by Fernández Dobao’s (2002) study. The final research question investigates how L1 Chinese may influence the acquisition of L2. If the knowledge of L1 does get transferred into L2, the result will support Ellis (1997) that L1 transfer is a cognitive process and positively influences L2 learning.. 9.

(21) 1.4 Significance of the Study The present study focuses on Chinese learners of Japanese kanji by examining whether the participants would be influenced by their L1 knowledge. Since previous studies like Haththotuwa Gamage (2003) tend to put more emphasis on the alphabetic learners of Japanese, in this thesis the participants are different from those in the previous research. In addition, a new classification of kanji will be proposed to substitute for the classifications of the previous studies (Kess & Miyamoto, 1999; Toyoda, 2009). Moreover, we will examine how form and meaning interact during the process of acquisition. The pictorial cues from which participants can infer meanings of kanji characters are also include in the task design of present study.. 1.5 Organization of the Thesis This thesis is organized as follows. Chapter Two reviews some theoretical and empirical studies on Japanese kanji, such as processing, previous classifications and some issues of form and meaning. A new classification of kanji is provided in Chapter Three, together with the research design of the present study. Chapter Four presents the overall results, and the discussion is also included. Chapter Five concludes this thesis.. 10.

(22) CHAPTER TWO LITERATURE REVIEW In this chapter, theoretical and empirical studies on Japanese kanji characters are discussed. Section 2.1 reviews three theoretical studies on the processing of kanji characters and introduces the classifications of kanji. Section 2.2 reviews four empirical studies on the recognition of kanji characters from the aspects of semantic and phonological and on the influence of L1. The context and word morphology are also discussed. Finally, a summary of this chapter is presented in Section 2.3. 2.1 Previous Theoretical Studies of Japanese Kana and Kanji In the literatures, kanji characters have been discussed in different perspectives and this section provides a general idea about kanji characters. Leong and Tamaoka (1998) discussed the processing of Japanese kana and kanji characters. Kess and Miyamoto (1999) introduced the origination of kanji characters and provided a classification of these kanji characters based on the internal structure. They also discussed the processing differences between kana and kanji. In addition, Nagano and Shimada (2014) examined the dual pronunciations of kanji characters from the perspective of the lexeme-based theory.. 2.1.1 Leong and Tamaoka (1998) From the cognitive processing perspective, Leong and Tamaoka (1998) were interested in finding the similarities and differences between different languages and how lexical items helped to process these languages. As an introduction to the Chinese and Japanese cognitive processing, the processing structures of the two languages were discussed. Two main research issues were addressed. First, Chinese as a morphosyllabic language, it would be interesting to see whether its phonological information would facilitate for analytical word 11.

(23) reading or not. Although Chinese speakers might not be able to process the phonological segmental information, such as phonemes and morphophonemes, Leong and Tamaoka still wondered to what extent word recognition and naming in Chinese would be aided by phonological representation. Additionally, there are two writing systems in Japanese, kanji and kana. Each of these two systems involves two aspects. As for kanji characters, its processing involves multiple readings (kun-reading and on-reading1) and representations of meanings. Kana with moraic segments represents timing and subsyllabic units. Hence, the processing of Japanese may then consist of two different cognitive structures. As mentioned by Leong and Tamaoka, Tan and Perfetti (1998) tested the Identification-with-phonology Hypothesis 2 with Chinese single-kanji characters and two-kanji characters and the converging priming was used. Their findings showed that phonology was an important factor in Chinese word recognition. When graphic recognition was activated, phonology information was activated simultaneously. Besides the time order between graphic information and phonological information, the latter was activated before semantics, and semantics might be influenced by phonology. Other studies investigated the Chinese homophones. From the study of Weekes, Chen and Lin (n.d.3), homophones facilitated the recognition of Chinese compounds. While the semantic priming effect was activated for compounds, the phonological priming effect only occurred in certain compounds of which characters consist of separate radicals. Those characters without separate radicals were found unable to facilitate phonological priming. Weekes et al. (n.d.) claimed that this difference in facilitation of homophone primes was due to the competition between different phonological forms for the correct pronunciation of a compound character. The semantic priming effect, however, was 1. These terms have been introduced in Footnote 2 on p.5. The Identification-with-phonology Hypothesis postulates that the phonology aspects are processed earlier than semantic aspects in recognizing (Chinese) characters. 3 The n.d. here means ‘no date’. The mentioned studies were not marked date in Leong and Tamaoka (1998), and hence the years of these studies were marked n.d. in this thesis. 2. 12.

(24) considered to be the semantic nature of Chinese orthography. In addition, Leong and Tamaoka mentioned that about seventy percentages of kanji were found two-kanji characters4, which could be divided into two types. According to the concept of each character which could be seen as a morpheme, the first type consisted of two similar concepts of the kanji morphemes and the second type consisted of two opposite concepts. The processing of these two types of two-kanji characters was affected by the activation of semantic representations at the morpheme level (Tamaoka & Hatsuzuka, n.d.). When processing of the two types of two-kanji characters was compared, it was found that two-kanji characters of the second type were processed slower, showing that the semantic representations of the second type competed at the concept level and hence slowed down the lexical decision. Besides the semantic representation, Leong and Tamaoka found phonological information was commonly used in processing Japanese kanji. Some radicals of kanji characters could convey meaning while some components of kanji could represent a speech-sound relation. Therefore, phonology, semantic and orthography were found important for processing kanji. With regard to on-reading and kun-reading in Japanese, Wydell (n.d.) argued that the computation of kanji was influenced by the structural differences between on-reading and kun-reading. According to Hirose (n.d.), Japanese people differentiate the two readings by the degree of a one-kanji semantic independence. Moreover, Yamada (n.d.) found that kanji and kana were processed differently. Words written in form of hiragana were processed faster than kanji characters which actually the same words as those hiragana words. However, when the participants were asked to translate these words into English, kanji characters were translated faster than hiragana words. Thus, Yamada argued that kanji characters were more closely related to meaning, that is, kanji might be semantic in nature. In addition, the participants with different 4. In this thesis, the kanji compounds are called two-kanji characters for consistency. 13.

(25) background knowledge of characters made errors in different aspects. While the native controls Japanese made more errors with kanji phonology, alphabetic background learners made orthographic error more often. Although both groups made errors with phonology or orthography, the errors were not with semantics, implying that kanji is more related with semantics. In sum, Leong and Tamaoka’s study provided an overview of previous research. Although the phonological aspect of kanji characters was usually considered less important, the phonological processing was actually working. While kanji characters were found more related to semantics, the phonological aspect still appeared to play a role in the Japanese processing.. 2.1.2 Kess and Miyamoto (1999) With regard to history and the internal structures of kanji characters, Kess and Miyamoto reported some background knowledge of kanji. They defined kanji characters as the preserved characters when kana is defined as a simplified form of these borrowed characters. For example, the Chinese character 安 is simplified into the hiragana symbol あ /a/ and the character 安 is still preserved as a kanji element in the Japanese vocabulary, such as 安い ‘cheap’. Besides the preserved kanji characters and transformed kana words, Japanese create their own kanji characters. They borrow components of Chinese characters and reorganize them, such as 辻 ‘cross-roads’ which is a combination of Chinese characters 十 and 辶. These newly formed characters cannot be found in Chinese5. In recent Japanese, although there are a large number of Chinese characters introduced into Japanese and many kanji characters are created, there are not many daily used kanji 5. Actually, Chinese is not the only source which forms the Japanese language. Japanese people also borrowed words from some other languages. For example, they used English word ‘hotel’ as ホテル /hoteru/. These words borrowed from other languages were usually written in katakana. Hence, Japanese is so unique that it contains three kinds of lexical items, hiragana, katakana and kanji (preserved kanji and newly created kanji characters). 14.

(26) characters. Among 50,000 possible kanji, there are only 3,000 kanji characters used in the magazine and newspaper. The used kanji characters gradually decrease as the time goes by. While there were 39.3% kanji characters used in novels written in 1900, but only 27.5 % kanji characters were used in novels written in 1950 (Kess & Miyamoto, 1999, p.21). However, kanji characters are still important in Japanese, since they are usually the essence of a sentence (i.e. the content words). Kanji characters can also be grouped into four types according to their structure and usage. First, the pictographic kanji describes the shape of objects. For example, the 口 ‘mounth’ describes the round shape of mouth. Second, the diagrammatical kanji can be called indicative abstract concepts, such as the number 二 ‘two’ or spatial relation 下 ‘down’. The third type is the compound-semantic kanji, which composed of two simple characters. Its meaning is the combination of the meanings of two characters. For example, the kanji character 鳴(る) is the combination of 口 ‘mouth’ and 鳥 ‘bird’ and means ‘twitter’. The last type is the phonetic-semantic kanji and it forms the majority of the kanji characters. This type of kanji characters consist of two radicals, the phonetic radical usually on the right-hand of the characters and the semantic radical usually on the left-hand of the characters. The phonetic radical denotes the pronunciation of kanji characters while the semantic radical conveys a general meaning of the kanji characters. For example, kanji characters like 偉 ‘great’ and 緯 ‘latitude’, include same phonetic radical 韋 on the right-hand side and both of these two-kanji characters are pronounced as /i/. However, kanji characters like 机 /tsukue/ ‘desk’ and 柵 /saku/ ‘fence’ are pronounced differently but they include the same semantic radical 木 and their meanings are related to ‘wood’. According to the structural differences of these phonetic-semantic kanji characters, the radicals in these characters can be at different positions. All of these different radicals are. 15.

(27) listed as follows, including the left-hand radical and right-hand radical mentioned above. (1). a. hen (right-hand). e.g. 波 ‘wave’ and 河 ‘river’. b. tsukuri (left-hand). e.g. 清 ‘clear water’ and 晴 ‘clear weather’. c. kanmuri (top). e.g. 草 ‘grass’ and 花 ‘flower’. d. ashi (bottom). e.g. 熊 ‘bear’ and 馬 ‘horse’. e. kamae (enclosure). e.g. 国 ‘country’ and 囲 ‘to enclose’. f. tare (top-left). e.g. 庭 ‘garden’ and 店 ‘shop’. g. nyoo (left-bottom). e.g. 返 ‘to return’ and 逃 ‘to escape’ (Kess & Miyamoto, 1999, p.36). With these different graphic configurations, a kanji character can be simple as well as complex. For example, 金 /kane/ can be alone as a word meaning ‘money’ or ‘metal’. It can also be a component of a more complex kanji character, such as 釘 ‘nail’ and 鉄 ‘iron’. However, there are some radicals which cannot stand alone. When the character 水 ‘water’ becomes a radical of a complex kanji character, the shape will be different as 氵. The left radical of 海 ‘ocean’ and 河 ‘river’ are the examples of which component 氵 denotes ‘water’. However, it cannot stand alone. Japanese multiple reading is an interesting issue as well. On the one hand, a kanji character can include multiple on-readings and kun-reaings6. For example, the kanji character 頭 can be read as /tou/ and /zu/ in multiple on-readings and as /atama/ and /kashira/ in the multiple kun-readings. On the other hand, a given kun-reading can represent multiple kanji characters. For example, /ki/ can represent 木, 季, and 生 etc. Although these different kanji characters with the same pronunciation may cause difficulty in memorizing, there is advantage for using these kanji homophones. A popular tongue twister kisha ga kisha de kisha shita ‘The reporter came back to his firm by train.’ is the best example. If we only have the pronunciation in hiragana, the meaning of this sentence may be ambiguous. When we 6. For these terms, please refer to Footnote 2 on p.5. 16.

(28) translate the popular tongue twister into kanji characters, the meaning becomes clearer as. 記者が汽車で帰社した7 ‘The reporter (記者) came back to his firm (帰社) by train (汽車).’ As for the processing of kanji, Kess and Miyamoto discussed the lexical access of mental processing. Since the Japanese writing system is complicated, the mental lexicon may develop different ways of kanji processing in syllabic kana and logographic kanji characters. It has usually been claimed that the phonological-based kana would involve phonological decoding and retrieve meaning later while the morphological-based kanji characters would access to meaning directly without the mediation of phonology. As mentioned above, the majority of kanji characters are phonetic-semantic kanji. This kind of kanji characters can also be called as phonographs. The feature of phonetic-semantic kanji characters is that the two radicals represent the meanings and pronunciations, respectively. When the function of the semantic radical is agreed that it can be used to infer meaning, the function of the right-hand radical8 is concerned as well. Hirose (1992) found that kanji characters with the same right-hand radicals showed faster reaction times, and argued that the information of the right-hand radicals (i.e. the phonological information) was useful for Japanese readers to process kanji. Besides the phonological activation, Wang (1988) found that the time relation between phonological information processing and semantic information processing was manifested in a parallel mode. With regard to the two-kanji characters, Kess and Miyamoto argued that Japanese readers might have an inventory of kanji building blocks. They used this inventory to cooperate with world knowledge and contextual information and helped readers to infer the. 7. Sometimes, different readings may compete with each other. However, it was found that for some kanji characters their kun-readings were prior to their on-readings, such as 島 /shima/(kun-reading) which prior to 島 /tou/ (on-reading). For such a kanji characters, the frequency ratio between the two readings might be a factor. 8 There are only a small number of kanji characters which have reliable phonetic radicals in Japanese. However, these phonetic radicals still play an important role in processing. 17.

(29) meanings of kanji characters. Even though the two-kanji characters might be unknown words, the mental lexicon could help learners. This inventory might be a built-in conceptual categorization system provided by semantics. The process of the two-kanji characters was formed on the word level (two-kanji characters) rather than at a morpheme level (one-kanji character). This was confirmed by some previous studies (e.g. Morton, Sasanuma, Patterson, & Sakuma, 1992). It was found that readers who read mixed kana-kanji script spent less time than read kana script alone in Kess and Miyamoto’s study. Kaitao (1960) asked his participants to fill in the blanks with appropriate words. The mixed texts did facilitate more accurate responses than the hiragana-only text and the mixed text was also processed faster than the kanji-only text. Another study obtained a similar result that the nursery tales containing both kana and kanji scripts were read faster than the scientific essays usually written in kanji. Hence, according to Kess and Miyamoto, the difference among kana-only texts, kanji-only texts, and kana-kanji mixed texts might be due the familiarity of script. To summarize, kanji characters derived from Chinese characters are important in Japanese because they are content words of sentences and they help differentiate ambiguous sentences. The internal structures of kanji characters can also help learners indicate meanings. Since different radicals convey different information, the information is processed differently. The familiarity may also be a factor affecting the process of kanji characters.. 2.1.3 Nagano and Shimada (2014) Nagano and Shimada (2014) investigated the nature of kanji characters from a morphological viewpoint and argued that kanji orthography could reflect the working of lexeme-based morphology in Japanese grammar. The difficulty of learning Japanese results from the fact that Japanese simultaneously includes kana and kanji scripts. A learner of 18.

(30) Japanese needs to decide which script should be used in a given lexeme. For example, a word meaning ‘mountain’ is pronounced as /yama/ and it can be written as 山 (as a kanji character) or やま (as a kana). The difficulty of learning kanji not only lies in the orthography system but also the in dual pronunciation of a kanji character, the on-reading and kun-reading. The relation between on-reading and kun-reading in a given kanji is not one-to-one but many-to-many. Sometimes, a given on-reading can correspond to different kanji characters, that is, these kanji characters are homophones. For example, an on-reading /ken/ could represent several kanji characters, including 建, 犬, and 堅. Sometimes, a kun-reading can also correspond to several on-readings. The kanji character 見 is read as /miru/ in a kun-reading but it can both be read as /kan/ and /ken/ in an on-reading. Although these different readings may cause difficulties in acquiring Japanese kanji characters, it seems not that challenging for native Japanese speakers. Kanji characters are actively used in their daily life. Thus, Nagano and Shimada claimed that Japanese kanji characters should be viewed in terms of lexeme-based morphology. The lexeme-based theory considers sounds and meanings as separate systems and explores the working of an autonomous level mediating between the two. A lexeme projects a paradigm which contains a set of stems. These stems are morphemes and function to map between various morphosyntactic properties and various phonological word forms. Hence, in this theory, a kanji character can be seen as a lexeme and its dual pronunciations are phonologically unrelated stems as suppletions. The reason why a kanji character is seen as a lexeme is that when a grammatical context is given, the dual pronunciations of a kanji graph disappear, as in (2): (2). a. Inflected adverbial: 見 /mi-ø/ b. Inflected hypothetical: 見 /mi-re/. When kanji character 見 used in the derivation form as in (3) and two-kanji characters formation in a kun-reading as in (4), the realization rule chooses one of the inflected forms as 19.

(31) in (2) to be the stem. For example, if (2a), the inflected adverbial form /mi-ø/, is chosen, the derivational form of agentive, derivational form of action noun and two-kanji characters formation in a kun-reading will be as in (3)-(4). (3). a. Derivational agentive with a suffix –te: 見 /mi-te/ b. Derivational action noun: 見 /mi/. (4) Two-kanji characters formation in a kun-reading: 見所 /mi-dokoro/ ‘merit’ As for the two-kanji characters forming in an on-reading, the inflected form /mi-ø/ cannot be used, but a more formal kanji reading which is read as /ken/ is used. This difference in the formation of two-kanji characters is not related to any semantic or phonology aspects but involves morphological [± bound] and pragmatic [± formal] features. While the derivation and two-kanji characters in a kun-reading involve [-bound] and [-formal] features, two-kanji characters in an on-reading involve [+formal] and [+bound] features. These two features define the paradigm of Japanese verbs and nouns. For example, the two-kanji characters 山道 ‘road of mountain’ is realized as /yama-michi/ in a kun-reading but /san-doo/ in an on-reading. Both readings denote the same meanings but they convey different degrees of formalness and boudness. As discussed by Nagano and Shimada, the kanji suppletion is developmental in history. The origin of dual pronunciations is from the language contact between Japanese and Chinese but the contact of two languages generates synonyms9. To sum up, kanji characters are complex that their dual pronunciations also cause difficulties in acquiring kanji. According to the lexeme-based theory, the two-readings of a kanji character is seen as a suppletion of the kanji and the development of suppletion is due to the avoidance of synonyms.. 9. This would violate the universal principles of iconicity among lexemes, especially the principles of lexemic antisynonymy and lexemic antihomophony. To avoid this situation, the solution is to use the avoidance strategy which could be done by means of morphology, that is, kanji suppletion. By mapping the synonymous native Japanese-Chinese pair to the different stems of a single lexical paradigm, the avoidance of synonyms and homonyms can be achieved. 20.

(32) 2.1.4 Summary In these studies, kanji is viewed from different perspectives. In terms of the internal structure of kanji characters, kanji characters can be divided into four types: the pictographic kanji, diagrammatical kanji and compound-semantic kanji. In the phonetic-semantic kanji, there are two radicals, the phonetic radical and the semantic radical. Based on the spatial relation, the radicals in kanji characters can be put in different positions. As listed, the radicals can be hen (right-hand), tsukuri (left-hand), kanmuri (top), ashi (bottom), kamae (enclosure), tare (top-left) and nyoo (left-bottom). These radicals not only can simply stand alone but also can combine with the other radical to form a kanji character (Kess & Miyamoto, 1999). With regard to the processing of kanji, the phonological processing and semantic processing have been examined. The kanji characters were found more related to a semantic aspect which seems to be the nature of kanji characters (Leong & Tamaoka 1998). Some radicals in kanji characters were found to denote a speech-sound relation, such as phonetic radicals and the time relation of the two types of processing was claimed to be in a parallel mode, in other words, the phonological information and semantic information were parallel processed (Kess & Miyamoto, 1999). Therefore, on kanji processing, both of semantic and phonology are factors which cannot be ignored. In addition, kanji characters can be viewed from a morphological perspective. As mentioned above, the processing of a two-kanji character which includes two separate concepts kanji characters would be slowed down. Hence, it was claimed that the characters in a two-kanji character were morphemes and the computation of characters was at the morpheme level (Leong & Tamaoka, 1998; Kess & Miyamoto, 1999). These kanji morphemes could stand as a lexeme. It is well-known that kanji is complex because of its dual pronunciations, that is, the on-readings and the kun-readings. This complex reading system did cause difficulties for native Japanese speakers. According to Nagano and 21.

(33) Shimada (2014), the two readings are suppletions of kanji characters. Thus, Japanese native speakers may view the suppletion as an innate mechanism. If L2 learners of Japanese are as good as natives in acquiring the two readings of kanji, we can explain that these learners also have a suppletion mechanism as native Japanese speakers do.. 2.2 Previous Empirical Studies of Kanji This section reviews some previous empirical studies of Japanese kanji characters and two-kanji characters. Learning strategies and influential factors like context are discussed. Chikamatsu (1996) investigated the L1 transfer effect from two language backgrounds and Mori (2003) examined how contextual clues and word morphology helped learners judge meanings. Haththotuwa Gamage (2003) explored learner strategies used to process kanji. Toyoda (2009) conducted a study on the development of learners’ awareness of kanji structures and readings by learners at different proficiency levels.. 2.2.1 Chikamatsu (1996) Chikamatsu (1996) investigated the influence of first language orthography system on second language word recognition. The issue focused on how the equally new or familiar Japanese words recognized by English and Chinese learners. Chikamatsu aimed to know with distinct backgrounds, how English speakers and Chinese speakers were different in word recognition. One issue discussed in her study was the degree of phonological decoding involved in the recognition. Several researchers proposed an orthographic depth hypothesis (Frost, Katz & Bentin, 1987; Frost, 1994; Katz & Feldman, 1983; Tabossi & Laghi, 1992), which has been proposed to illustrate the degree of phonological coding in reading strategies. When a language involves more phonological information, the language is called as a “shallow” language. In contrast, if a language utilizes less phonological coding, the language is a “deep” language. Although Chinese is a deep language, it has been argued that 22.

(34) phonological information is still activated in Chinese and the phonological information is primarily automatic involved in word recognition. However, the degree of phonological involvement was still different from English speakers. This difference of involvement may be due to the orthographic effect which is an important factor for word recognition. In Chikamatsu’s study, two hypotheses were testified: (1) the type of orthographic system would influence the word recognition strategies (the visual coding or phonological coding), and (2) the first language transfer would occur that L1 orthographic word recognition affected on second language word recognition. Thus, for logographic Chinese speakers, it was predicted that they would utilize more visual coding while syllabic English learners would use more phonological coding. The participants were 45 Americans and 17 Chinese. Both groups of the participants were college students and took regular second language courses at school. A kana lexical judgment test was conducted to the participants’ word recognition strategies. All test items were written in kana. In this test, (visual) familiarity and word length factors were examined. Hence, according to familiarity, there were three conditions, visually familiar, visually unfamiliar, and non-word. The word length was designed from two to five kana letters. Two procedures were implemented in the study. The participants’ Japanese language skills were first tested as a pretest by asking the participants to do the kana test and the vocabulary test. A lexical judgment test was also employed to judge the participants’ recognition of the test items which consisted of both Japanese and non-Japanese words. The performance of the participants was compared based on their reaction time (RT). It was predicted that the Chinese and English speakers’ reaction time on unfamiliar words would be longer than familiar words. However, for the English speakers, the rate of slowing down might be lower than the Chinese group. Comparing the unfamiliar words and non-words, it was predicted that both the Chinese and English groups would spend more time on non-words than on unfamiliar words. The rate of slowing down for the English 23.

(35) speakers was higher than the Chinese group. Furthermore, in the aspect of word length, the English participants’ reaction time was longer than the Chinese participants. The longer the words were, the more time the English group would need to convert kana into sound units. The Chinese speakers, on the contrary, considered these letters as a whole. The results showed that the reaction time did gradually increase for both the Chinese and English groups when unfamiliar and non-words appeared. The English speakers’ rate of slowing down was less than that of the Chinese group in unfamiliar words but the difference was not significant in non-words. Longer words were identified slower than shorter words. The script effect was found as well. The hiragana words were processed faster than katakana words. In the aspect of word length, both of English and Chinese reaction time gradually increased when word length became longer but the English speakers did show a higher reaction time. The Chinese participants performed a higher rate of slowing on the reaction time than the English participants when visually unfamiliar words appeared. This showed that the Chinese participants did rely more on the visual information, and the English speakers utilized the phonological information more. Nevertheless, the insignificant result for the non-words meant that the participants slowed down at a similar rate. Concerning the word length effect, the slower performance of the English participants on longer words, especially in the hiragana condition, showed that the word length and the script effect were influential. To sum up, the Chinese learners and English learners were found to perform differently on the word recognition test. All of the results showed that while the Chinese participants applied more visual information from their L1, and that the English participants used more phonological information. Thus, both of participants transferred their L1 orthographic knowledge into L2 Japanese word recognition.. 24.

(36) 2.2.2 Mori (2003) Mori (2003) tested the interaction between contextual clues and morphological information. The contextual clues and word morphology were important information because they could help learners infer meanings of the novel words. The word morphology could provide semantic information as well as phonetic information, such as the phonetic radicals and semantic radicals. However, this information might not always reliable in inferring meaning of Japanese two-kanji characters. An example mentioned in Mori’s study is the two-kanji characters, 外見 ‘appearance’. Although the learners might be familiar with the one-kanji character of 外 ‘outside’ or 見 ‘to see’, they still incorrectly inferred the meaning of the two-kanji characters. This kind of two-kanji characters is well-known on their semantic semitransparency. The semantic semitransparency of two-kanji characters is a combination of two familiar characters with high frequency but meanings cannot be inferred from combining the two meanings of characters. Hence, other than the word morphology, learners sometimes need to find other clues in order to obtain meanings. One method is to find information from the context. Thus, the purpose of Moris’ study was to find whether the contextual information was as informative as word morphology or not. Seventy-four English college students were recruited in Mori’s study. They had learned Japanese for at least two semesters, and they were asked to complete two tests, the kanji pretest and the two-kanji character test. The kanji pretest was designed to test the learners’ Japanese kanji characters knowledge, where the participants had to write down the information of the characters. The second test, two-kanji character test, consisted of 45 test items with semantic semitransparent characters, which were further assigned to three sets. Each set represented a condition. The first condition was the kanji-characters-only condition (hereafter the K condition) in which the two-kanji characters were presented in isolation and the participants were required to write down the meanings of two-kanji characters in English. The second was the context-only condition (hereafter the C condition), where the 25.

(37) sentence was presented without kanji characters and the participants had to fill in the blanks with possible words. The third condition was both-kanji-and-context condition (hereafter the KC condition), where the two-kanji characters were presented in the sentence and the participants had to write down possible meanings in English. Four variables were also examined, semantic relatedness between students’ best guesses and meaning, students’ confidence, semantic relatedness among an individual’s guesses, and agreement of part of speech. As for the first variable, semantic relatedness between students’ best guesses and meaning, the result of the KC condition was significant and showed higher means than the other two conditions. The participants provided better guesses by combining the kanji characters and the context clues. There seemed no significant difference between the K condition and the C condition. Moreover, the result for the students’ confidence variable was that the score for the K condition was lower, showing that the participants had more confidence in their guesses when the contexts were presented. This indicated that sentences containing more familiar elements like kana words were easies to process. In the third semantic relatedness among an individual’s guesses variable, the mean score for the KC condition was the highest and followed by the K condition. The C condition received the lowest score, showing that the students could generate more semantically related guesses when kanji characters and context information were provided. Finally, it was found that the score for the C condition was the highest while the K condition received the lowest score. The difference between the C condition and the KC conditions was not statistically significant but the difference between the C condition and the K condition was significant. This implied that the availability of contextual information would influence the result of syntactical related guesses. In other words, if contextual cues were available, the participants would generate more syntactically related answers. For 26.

(38) example, a Japanese grammatical category was indicated by affixes in hiragana, such as する (verb indicator). Therefore, the immediate context helped the participants determine the part of speech for the word. However, the contextual clues sometimes could be misleading because more than one answer would be possible and cause the difficulty. In sum, the word morphology could help infer the meaning of kanji and context information could help indicate the syntactic categories. When the context was offered, the participants would also have more confidence in their test judgment.. 2.2.3 Haththotuwa Gamage (2003) As shown in Chikamatsu’s (1996) study, the English and Chinese learners employed different kanji-word recognition strategies. For example, the English speakers utilized the phonological coding strategies while the Chinese speakers utilized the visual coding strategies. Haththotuwa Gamage (2003) conducted a task on kanji learning strategies to find whether there were differences in strategy uses between alphabetic background learners and character background learners and which strategies were more efficient. The participants participated in this research were 116 sophomore, including 64 alphabetic background learners and 52 character background learners. They had learnt about 120 kanji. A questionnaire was employed in this study. Some previous researchers had identified three types of strategies to study the kanji processing, that is, shape (as visual strategies), pronunciation (as phonological strategies) and the meaning (as semantic strategy) of kanji. In this study, Haththotuwa Gamage incorporated these three types of strategies in her research questionnaire. The participants had to read statements. There were totally 20 statements and each statement represented a strategy which could be used in learning kanji. Based on the participants’ previous learning experience, they had to indicate their frequency of use and perceived efficiency of the strategies from the 20 statements. Regarding the perceived efficacy of learners, the result showed that the mean 27.