SCIENCE REGISTER ACROSS CHINESE AND ENGLISH: THE RELATION BETWEEN LEARNERS' LANGUAGE PRODUCTION OF SCIENTIFIC CLASSIFICATION DISCOURSE IN CHINESE AND ENGLISH by ESTHER KA-MAN TONG B.A. (Honors), City University of Hong Kong, 1996 Dip. Ed., The University of British Columbia, 1998 A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF ARTS in THE FACULTY OF GRADUATE STUDIES (Department of Language and Literacy Education; Faculty of Education) We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA August 2004 © Esther Ka-Man Tong, 2004 THE UNIVERSITY OF BRITISH C O L U M B I A FACULTY OF GRADUATE STUDIES Library Authorization In present ing this thes is in partial ful f i l lment o f the requ i rements for a n advanced degree at the Universi ty of Brit ish Co lumbia , I agree that the Library shall m a k e it f reely avai lable for reference and study. I fur ther agree that permiss ion for ex tens ive copy ing o f this thesis for scholar ly pu rposes m a y be gran ted by the head of my depar tmen t or by his or her representat ives. It is unders tood tha t copy ing or publ icat ion of this thes is for f inancial ga in shall not be a l lowed wi thout my wr i t ten permiss ion. T O N G , Esther Ka M a n 03/09/2004 N a m e of Au thor (please print) Date (dd/mm/yyyy) Tit le of Thes is : Sc ience register ac ross Ch inese and Engl ish: The relation be tween learners ' language product ion of scienti f ic classi f icat ion d iscourse in Ch inese a n d Engl ish Degree: Master of Ar ts in Literacy Educat ion (TESL) Year: Depar tment o f Language and Li teracy Educat ion (LLED) T h e Universi ty of Brit ish Co lumbia Vancouver , BC C a n a d a rad.ubc.ca/forms/?formlD=THS page 1 of 1 last updated: 1-Sep-04 ABSTRACT Faced with a large inflow of students with limited English proficiency, educators in Canada are being challenged to prepare these students to learn content subjects in the mainstream classes through their second language, English. To help these language minority students achieve academically, Mohan (1986) addressed the need for students to have knowledge of genres, and in particular, an understanding of the knowledge structures involved in content areas to construe the content of their course. Classification, together with five other semantic structures of discourse which Mohan has identified, can be captured visually and linguistically constructed using specific language features. Tang (1994b) noted that similar visual representations of knowledge structures occur across languages. These views line up with Halliday (1993), who suggested that there are common features of science registers in English and Chinese. The current resarch explores these notions in detail, concentrating on the knowledge structure of classification. Specifically, this paper focused on exploring how the participants' first and second languages, Chinese and English, were used to create meaning and understanding of knowledge acquired through the participants' first language, Chinese. The analysis concentrated on identifying the ideational meaning of the texts collected in tasks which involved classifying living things on Earth, and then looking into how the participants used their LI and L2 to represent the classification knowledge structure in the time series design. The results showed that the participants were able to use both their LI and L2 as resources for learning and representing their knowledge of classification. Their learning of the new classification knowledge became evident in both their LI and L2 texts, revealing that the content knowledge they acquired in their L I , Chinese was transferable to their L2, English. Knowledge structure graphics (KS graphics), which are graphic representation of the knowledge structures (Mohan, 1986), played the role of mediator, connecting the learners' background knowledge and the new knowledge. With the help of graphic organizers displaying the semantic structure and text structures of classification, the participants were able to use appropriate language features distinctive to the discourse of classification to ii realize their learning of the content knowledge in both their stronger language, Chinese, and weaker language, English. In addition, the comparison of the participants' post-reading classification texts, which were written based on their learning about the classification discourse in Chinese, and the texts produced after a one-hour instructional session about writing classification discourse in Chinese showed that instruction in the LI helped students write a more developed classification discourse not only in their LI but also in their L2. The participants were able to create a more extended discourse with an increasing number and a greater variety of language features used to realize the knowledge structure of classification. These findings suggest an interdependent relation between bilingual learners' language production of scientific classification discourse across their first and second languages, reinforcing Halliday's (1993, 1999) notion of a common language of science. Recognizing the important roles that KS graphics play in assisting the transfer and reconstruction of knowledge across languages, this study addresses the importance of extending learners' academic language learning experience in both their first and second language. TABLE OF CONTENTS Abstract List of tables vii List of figures ix Acknowledgements x C H A P T E R 1: S T A T E M E N T O F T H E P R O B L E M 1.1 Background to the study 1 1.2 The research questions 3 1.3 The rationale for the choice of research topic 4 1.4 The format of the thesis 5 C H A P T E R 2: R E V I E W O F T H E L I T E R A T U R E 2.1 The'culture'of science 6 2.2 Science discourse constructs 'categorizing, taxonomic organization 7 2.3 Developing scientific literacy 7 2.4 Scientific literacy across languages 8 2.5 The use of graphics in education 9 2.6 The knowledge framework and knowledge structure graphics 11 2.7 Summary of the chapter 16 C H A P T E R 3: R E S E A R C H M E T H O D O L O G Y 3.1 An overview of the research methodology 17 3.2 Design for data collection 17 3.2.1 The setting 18 3.2.2 The participants 18 3.2.3 The data collection 19 3.3 Systemic functional analysis of discourse 20 3.4 Summary of the discourse analysis procedures 23 3.5 The research questions 24 iv CHAPTER 4: THE DATA ANALYSIS 4.1 The primary research question 26 4.2 The analysis of the language features in the model English and Chinese classification texts 26 4.3 Student-constructed classification trees in the pre-reading and the post-reading tasks 27 4.4 Examining the effects of reading a model classification text 30 4.4.1 A comparison of the pre-reading and post-reading classification discourse ... 31 4.4.1.1 Quantitative results: Descriptive 31 4.4.1.2 Quantitative results: Statistical 37 4.4.1.3 Qualitative results 37 4.4.1.3.1 Common sense versus scientific classification 40 4.4.1.3.2 Transfer of the content structure across LI and L2 40 4.4.1.3.3 LI versus L2 classification discourse 41 4.5 Analysis of the post-instruction classification trees 49 4.6 Examining the effects of instruction 50 4.6.1 A comparison of the post-reading and post-instruction classification discourse 51 4.6.1.1 Quantitative results: Descriptive 51 4.6.1.2 Quantitative results: Statistical 57 4.6.1.3 Qualitative results 57 4.7 Possible effects of transfer on students' language production 65 4.8 Summary of the findings 68 CHAPTER 5: 5.1 Discussion of findings 70 5.1.1 Cummins's threshold theory and common underlying proficiency 70 5.1.2 Halliday's conception of science registers across languages 70 5.1.3 The potential roles of classification graphics and knowledge structure graphics generally—A means to represent the bilingual's understanding of transfer across languages 71 v 5.1.3.1 Organize knowledge, and stimulate questioning and reflection on the content knowledge 71 5.1.3.2 Represent knowledge of classification and other knowledge structures 71 5.1.3.3 Stimulate negotiation of meaning, and help learners develop and reconstruct their understanding of classification and other knowledge structures 72 5.1.3.4 Facilitate the transfer of knowledge across the LI and L2 72 5.1.3.5 Assist the realization of knowledge structures in LI and L2 academic discourse 73 5.2 Implications 75 5.2.1 Cummins's "Linguistic Interdependent Principle" and goals of ESL programs 75 5.2.2 Halliday and raising students' awareness of the language features of classification in their LI and L2 75 5.2.3 Knowledge structure graphics and mediating between students' languages and multimedia 76 5.3 Future directions 78 References 79 Appendix 84 vi LIST OF TABLES Table Title Page Table 3.1 The participants 19 Table 4.1 Language(s) used in the student-constructed pre-reading and post-reading classification trees 28 Table 4.2 A summary of the total number of language features found in the pre-reading and post-reading Chinese classification discourse 31 Table 4.3 Results of the analysis of the participants' pre-reading and post-reading Chinese classification discourse 32 Table 4.4 A summary of the total number of language features found in the pre-reading and post-reading English classification discourse 34 Table 4.5 Results of the analysis of the participants' pre-reading and post-reading English classification discourse 35 Table 4.6 The test statistics: Comparison of pre-reading and post-reading Chinese and English classification texts 37 Table 4.7 Examples of SU's texts 38 Table 4.8 Examples from a participant's pre-reading Chinese classification discourse 41 Table 4.9 Examples from a participant's pre-reading English classification discourse 45 Table 4.10 A summary of the total number of language features found in the post-reading and post-instruction Chinese classification discourse 52 Table 4.11 Results of the analysis of the participants' post-reading and post-instruction Chinese classification discourse 53 Table 4.12 A summary of the total number of language features found in the post-reading and post-instruction English classification discourse 54 Table 4.13 Results of the analysis of the participants' post-reading and post-instruction English classification discourse 55 Table 4.14 The test statistics: Comparison of pre-reading and post-reading Chinese and English classification texts 57 Table 4.15 Examples of SU's texts 58 Table 4.16 Noun phrases which define generalized participants 62 vii Table Title Page Table 4.17 The use of linking verbs and descriptive words 63 Table 4.18 The use of linking verbs post-instruction 63 Table 4.19 The use of action verbs to describe behavior 64 Table 4.20 The influence of Chinese on English 66 Table 4.21 A summary of SU's language features across the three writing tasks 69 vi i i LIST OF FIGURES Figure Title Page Figure 2.1 Cummins's common underlying proficiency model of bilingualism 9 Figure 2.2 Form-meaning relations in the field of a social practice 13 Figure 3.1 The design of the present study 17 Figure 4.1 Example of a student-constructed classification tree in the pre-reading task (Graphic drawn by Participant I) 29 Figure 4.2 Example of a student-constructed classification tree in the post-reading task (Graphic co-constructed by Participant I and Participant J) 29 Figure 4.3 Example of a student-constructed classification tree in the post-reading task (Graphic co-constructed by Participant PA & MA in both Chinese and English in the post-reading task) 30 Figure 4.4 An example of the classification tree modified by Participant MA in the post-instruction task 50 Figure 5.1 Roles of classification graphics in aiding the transfer of knowledge across languages 74 Figure 5.2 Re-conceptualizing Cummins' Language Proficiency Model (1980, 1981) and Mohan's Knowledge Framework (1986) 74 i x ACKNOWLEDGMENTS I wish to express my sincere gratitude to Dr. Bernard Mohan for his untiring guidance throughout the course of this study. I am deeply impressed by his patience and enthusiasm in helping me overcome the difficulties encountered in the course of this study. Without his assistance and support, this project would not have been possible. I am also grateful to Mr. Hanson Lau, Mr. Bernard Chan, and Mr. Michael Yau of Hanson Lau Education Center for granting me the opportunity to conduct the study with the students in their school. Special gratitude is extended to Dr. Tammy Slater for giving me inspiring feedback on content and mechanics to make this a better thesis. My appreciation is also extended to the fourteen students of Hanson Lau Education Center from whom the data of this study were collected. Without their cooperation, this study could not have been completed. Finally, I wish to thank my parents, husband, and sisters for their love and spiritual support. x CHAPTER 1: STATEMENT OF THE PROBLEM 1.1 Background to the study In the last century, the thinking on the topic of bilingualism and intelligence has changed from a negative perspective to a positive one. Baker (1993) divided the last hundred years into three periods: The period of detrimental effects (up to 1960), the period of neutral effects, and the period of additive effects (from 1962). Early research, studies done prior to the 1960s, was motivated by the perception that two sets of linguistic abilities are separate and that each individual has only a limited amount of linguistic capacity available. Sharing this capacity between two languages would therefore lead to lower levels of proficiency in each language compared to unilingual speakers. This view held that bilingualism was a negative force in children's personal and academic development. Acknowledging the problems with this early research, most researchers now seem to have reached a general agreement that bilingualism can positively affect both intellectual and linguistic progress. A large number of studies after Peal and Lambert's study (1962) have reported that bilingual children exhibit a greater sensitivity to linguistic meanings and may be more flexible in their thinking than are monolingual children. The Additive Bilingual Enrichment Principle contends that bilingual children are adding a second language to their repertory at no cost to the development of their first language (Baker, 1993). There are still some studies, however (e.g. Wiley, 1997; Crawford, 1998; Lessow-Hurley, 2003), which have indicated that it is still a commonly held belief that the best way to learn a second language is through total immersion in the target language. Facing a large influx of non-English immigrants, educators in Vancouver have for a long time acknowledged the need to assist ESL learners to cope with their learning of academic knowledge through English (see, for example, Early, 1991; Early, Mohan, & Hooper, 1989). Academic English, these educators have realized, is more challenging for students than everyday English. Along this line, Cummins (1981) distinguished Basic Interpersonal Communicative Skills (BICS) with Cognitive Academic Language Proficiency (CALP). BICS refers to those aspects of language required for daily interpersonal 1 communication, whereas C A L P refers to those skills required for academic studies and serious higher-level thinking. Cummins claimed that these two levels of language proficiency proceed relatively independently (Cummins, 1980). BICS may be relatively quickly mastered in a second language situation —say, in a year or two—and may be developed relatively independent of one's first language. On the other hand, CALP, which is thought to be closely related to literacy skills and academic performance in schools (i.e., higher order cognitive processes such as synthesis, discussion, analysis, evaluation, and interpretation) and to one's skills in these areas in the first language, takes more than five years to acquire when a second language is used as the medium of instruction. Cummins's theory explains why ESL learners who are rated as fluent speakers of English may still encounter difficulties in learning academic subjects at school. Many language minority students are forced to study in the regular curriculum before they have fully mastered C A L P as educators commonly believe that "there is simply no time to delay academic instruction until these students have developed high levels of English language proficiency" (Short, 1993, p. 628). Therefore, there is a need to develop ways to expand these ESL students' higher order cognitive thinking skills and their knowledge of new content while developing their second language proficiency. One essential task that students have to cope with in their academic studies is to deal with science content. As Halliday (1993) acknowledged, "students of all ages may find them (scientific texts| hard to read, and we know from various research reports that, in English at least, the difficulty is largely a linguistic one" (p. 124). In this respect, it is important that more effective teaching methods be developed to help ESL learners acquire the language skills needed for understanding and expressing the semantic structure of scientific knowledge in their academic studies. Believing that maximizing ESL learners' exposure to English can accelerate their second language learning, many people have supported an "English Only Policy" in the classrooms (see Krashen, 1996, for a discussion of this). Moreover, some parents worry that their children's use of their native language may hinder them from learning English in the short period of time they have to meet the demands of the schools (see Prince & Lawrence, 2 1994; Shum, 2004; Siren, 1991, for relevant discussions). However, Halliday (1993) suggested that there are certain shared features in the way scientific texts are organized and realized in languages. Cummins (1981) also believed that first and second language skills are interdependent. Mohan (1990) recognized the contribution of Cummins's Common Underlying Proficiency Model of Bilingualism to the integrated language and content approach, addressing the role that the learner's LI knowledge plays in the acquisition of an L2. It is considered that the language resources and knowledge learners carry with them from their first language are opportunities for positive transfer in language learning. Drawing on these conceptions, this study aims to explore the relation between learners' language production of scientific classification discourse in Chinese and English from a sequence of tasks on the classification of living organisms. Specifically, it investigates how students use both their first language, Chinese, and their second language, English, as resources for learning and representing their knowledge of classification. The main focus is the nature of bilingual learner classification discourse in both languages, since little is known about this. The study will be given the form of a time-series design, but the main emphasis is on the "effects," the little-known area, rather than precise causal relations. That is, the main interest will focus on the comparison and contrast of student writing at six stages in a time series. Consequently the primary research question is the first one stated below; several others were asked so that this question could be responded to fully. 1.2 The research questions As stated above, the primary interest of this thesis revolves around the comparison of student writing at six stages in a time series which was evident in a unit on classification discourse taught to high school students whose first language was Chinese and their second, English. There is therefore one primary question. Primary research question: Given Halliday's view that there are commonalities of science register across Chinese and English with respect to scientific classification, how do students use Chinese and English to represent the knowledge structure of classification in the tasks studied? What are some of the similarities and differences? To help respond to the above research question, the following were asked: • How do the participants represent their initial knowledge of the classification in a visual format? Does the reading of a model text of classification change the students' visual representations? • Does reading a model text of classification of Chinese help learners to write more developed classification in Chinese? • Does reading a model text of classification of Chinese help learners to write more developed classification in English? • Does instruction in the discourse features of classification in Chinese change the students' visual representations? • Does instruction in the discourse features of classification in Chinese help learners to write more developed classification in Chinese? • Does instruction in the discourse features of classification in Chinese help learners to write more developed classification in English? Are there positive and/or negative effects of transfer on students' language production in the tasks? 1.3 The rationale for the choice of research topic As suggested at the beginning of this chapter, teaching language in isolation has been recognized as inadequate in preparing ESL students for content learning, and there has thus been a shift away from "isolating" to "integrating" language and content instruction. This approach aims to support ESL students' academic and cognitive development while they are in the process of acquiring English. To assist ESL learners' learning of content knowledge through English and language learning through content, Mohan (1986) proposed the "Knowledge Framework" to assist both language and content teachers in the systematic planning and monitoring of the integration of language and content instruction (Mohan, 1990). Graphic organizers, which are recognized as another means to represent the semantic structures of knowledge (Mohan, 1986), are claimed to be effective tools to aid students' comprehension of new knowledge (e.g., Andersson & Barnitz, 1984; Carrell, 1984; 1985; 1994) as well as the presentation of content knowledge in discourse (Mohan, 1989). 4 With this information about graphics and knowledge structures in mind, I designed an initial study to explore the gap between the "potential" and "actual" uses of graphics, and to investigate one ESL student's attitude toward the role of graphics in facilitating her learning of geography through English in Hong Kong and Canada (see Tong, 2000). This first examination indicated that using graphics to develop academic language—to promote and bridge the gap between content and language learning—was not one of the functions of the graphics promoted in the geography textbook I analyzed. Moreover, the role of graphics in the textbook I analyzed was limited to aiding students' content comprehension (the receptive use of graphics) instead of production, or the presentation of content in text. Most of the exercises in the activity book were restricted to those which demanded very little language production; for example, they involved labeling diagrams, matching pictures with isolated sentences, filling in the blanks of the sentences presented in a table form, and choosing the correct answer from different choices presented in a picture. As Tang (1994b) suggested, although similar forms of graphics found in the Hong Kong textbooks appear across languages, there is no guarantee that the different functions of graphics—for example, generative, representative/explanatory, and evaluative—are systemically and fully applied and developed by textbook writers, teachers, and students. Recognizing that graphics have the potential to be used to develop receptive and productive functions in real classroom practices, I designed the current study to integrate the use of knowledge structure graphics into the content and language learning tasks. I aimed to explore the productive functions of classification graphics in facilitating the transfer of thinking skills across learners' first and second language and in aiding students' language production of Chinese and English classification texts. 1.4 The format of the thesis Chapter 1 has briefly introduced the problem which this thesis will investigate and stated the questions which guided the exploration. Chapter 2 will review the literature which is relevant to the questions. Chapter 3 will describe the methodology which was used to respond to the questions, and Chapter 4 will present the findings in detail. Chapter 5 will conclude with a discussion of the issues and implications raised in this study. 5 CHAPTER 2: REVIEW OF THE LITERATURE 2.1 The 'culture' of science My research questions draw upon Halliday's conception of science registers across languages (Halliday & Martin, 1993; Halliday, 1999). Discussing common features of science registers in English and Chinese, Halliday asks, "is there in some sense a single 'language of science' that is essentially the same no matter what language it is manifested in? Of course English and Chinese use different expression systems and different sentence grammars. But it might be that their scientific registers would have many features in common" (Halliday & Martin, 1993, pp. 124-5). Halliday considers it likely "that nominalising metaphors in grammar are common to all languages which are used in the construction of science as currently understood in the late twentieth century" (Halliday & Martin, 1993, p. 53). Given that common features are found, further questions are: Where (at what levels) in the linguistic system do we explain them? Are they in the lexicogrammar —that is, in the meanings constructed into sentences by the syntax and the vocabulary? In the discourse—the composition of the text and its rhetorical structure? Or in some higher order—the ideological framework of knowledge, beliefs and value systems that form the cultural context of the text? ... the answer may be, of course, that the concept of scientific language involves all three" (Halliday & Martin, 1993, p. 124). Halliday uses context of culture in a distinctive sense which he illustrates concretely when discussing himself teaching a class of scientific Chinese to a group of anglophone Cambridge scientists: There certainly was a context of culture; and you couldn't hope to learn scientific Chinese without knowing quite a lot about it. But culture here does not mean the traditional culture of China. It means the culture of modern science, whether practised by Chinese or English or Australian or Vietnamese or any other nationality of scientists. When we talk of the cultural context for language education, we have to go beyond the popular notion of culture as something defined solely by one's ethnic origins. All of us participate in many simultaneous cultures... When we talk of the context of culture for language activities we mean those features of culture that are relevant to the register in question. (Halliday, 1999, pp. 17-18) It is this sense of culture that this thesis is exploring. 6 2.2 Science discourse constructs 'categorizing, taxonomic organization' Halliday has particularly noted two characteristics of science discourse: taxonomies and chains of reasoning. Science discourse constructs "categorising, taxonomic organisation" and "reasoning, logical progression" (Halliday, 1998, p. 201). The concentration here will be on the first, "categorising, taxonomic organisation." Unsworth (2001) explains: The language of subjects like science and geography are characterized by the development of technicality. That is, science reinterprets experience technically by defining elements of experience using technical terms (for example ecosystem, biome). These technical terms are ordered taxonomically. (p. 130) In other words, science reorganizes commonsense, everyday experience into 'uncommonsense' technical taxonomies. The term 'technical taxonomy' will be used to refer to an ordered, systematic classification that creates technical meanings. The term 'classification' will be used as a broad cover term that includes both commonsense classifications and uncommonsense technical taxonomies. The linguistic features central to classification and taxonomy have been discussed in detail in chapters eight through eleven of Halliday and Martin (1993), where it is also noted that diagrams which represent classification and taxonomy occur frequently in textbooks and that teachers can use similar diagrams to prepare students to write science (see Halliday & Martin, 1993, p. 174). 2.3 Developing scientific literacy Unsworth (2001) reviewed current work in Systemic Functional Linguistics for the development of literacy in the context of content area teaching. He described the Literacy Development Cycle (LDC) as a set of strategies for designing learning experiences in literacy. In its most generalized form the LDC incorporates modelled, guided, and independent practice in the comprehension and composition of texts. Guidelines for implementing the L D C with respect to curriculum area writing (p. 236) include (a) having the students read examples of the genre being studied, and (b) instructing students in identifying the purposes and stages of the genre and the significant grammatical features of the genre. 7 2.4 Scientific literacy across languages The above work in Systemic Functional Linguistics can be related to research in bilingualism. In his work on bilingual language proficiency, Cummins argues for "the Linguistic Interdependence Principle" (Cummins, 1987), meaning "that first and second language academic skills are interdependent," (p. 29), or in other words, manifestations of a common underlying proficiency. The interdependence principle has been stated formally as follows (Cummins, 1981): To the extent that instruction in Lx is effective in promoting proficiency in Lx, transfer of this proficiency to Ly will occur provided there is adequate exposure to Ly (either in school or environment) and adequate motivation to learn Ly. (p. 29) This interdependence principle addresses the importance of ESL learners' past experience and knowledge in their first language (LI) as it is believed that the thinking skills which ESL learners acquired through their LI may aid their understanding of the same concept presented in their L2. In concrete terms, what this principle means is that in, for example, a Japanese-English bilingual program in Japan, English instruction that develops English reading and writing skills is not just developing English skills, it is also developing a deeper conceptual and linguistic proficiency that contributes significantly to the development of literacy in the majority language (Japanese). In other words, although the surface aspects of different languages (e.g., pronunciation, fluency, etc.) are clearly separate, there is an underlying cognitive/academic proficiency that is common across languages. This "common underlying proficiency" makes possible the transfer of cognitive/academic or literacy-related skills across languages" (Cummins, 1980; see also Krashen, 1996, for a discussion of literacy transfer). Baker (1993) also suggests: Depending on language development in both languages, the cognitive functioning of an individual can be viewed as integrated, with easy transfer of concepts and knowledge between languages, (p. 145) Cummins's Iceberg Analogy theory (Cummins, 1980, 1981) explains the relationship between bilingualism and thinking processes and products. Both the LI and the L2 are claimed to operate through the same central processing system (what Baker calls the "Central Operating System"), which enables the transfer of knowledge and thinking skills, as shown in Figure 2.1. 8 Figure 2.1 Cummins's common underlying proficiency model of bilingualism Note that the exact details of this common underlying proficiency are left open, a situation which has not gone uncriticized. The S F L work on classification provides a wealth of specific linguistic analysis which offers important directions for future research on transfer and interdependence between a bilingual's languages. 2.5 The use of graphics in education As stated in Chapter 1, teaching language in isolation has been recognized as inadequate in helping E S L students learn content in North America and Canada, and there has therefore been a shift away from 'isolating' to 'integrating' language and content instruction, and approach with aims to support E S L students'academic and cognitive development while they are in the process of acquiring English. Because graphics are able to display the text structures (the formal schemata) of the content knowledge, they have been considered to be effective tools for making content learning through language or language learning through content more comprehensible and manageable to the second/foreign language learners, conditions which Krashen (1996) claims to be essential for second language acquisition to occur. The facilitative effect on reading comprehension is highlighted in previous research (e.g., Boothby & Alvermann, 1984; Griffin, Simmons, & Kameenui, 1991; Thompson, 1993). As Dunbar (1992) stated, 9 From a teaching perspective, a visual allows for language input through discussion of various ways of expressing what is shown in a visual, how ideas can be combined, how expression can be begin and end at various points on the visuals and allows students to test out and share with their classmates various forms of discourse. Students, therefore, supply a great deal of input for each other, (pp. 67-68) Nevertheless, other than facilitating reading comprehension and recall, other researchers (e.g., Mohan, 1990; Tong, 2000) have pointed out that graphic organizers can also be used to facilitate other aspects of learning such as writing, speaking, listening, and higher order thinking skills. In her case study which aimed to explore the "potential" and "actual" uses of graphics, Tong (2000) found that the role of using graphics productively to facilitate the students' presentation of content through language had often been neglected. Acknowledging the value of different functions of graphic organizers and the interdependent relationship between learner schema, content, language, knowledge structures (thinking skills), and graphics organizers, she suggested that all potentials (including the receptive and productive functions) of graphics should be explored in classroom practices. In other words, not only should graphics be used to aid students' comprehension of the texts, they should also be used to facilitate students' first and second language production. In addition, emphasizing the role of graphics in Integrated Language and Content Instruction (ILC), Tang has conducted several studies (1991a; 1991b; 1992a; 1992b; 1993; 1994a; 1994b; 1997) to examine the effect of graphics and to explore different aspects of using graphics in integrated language and content classes. The result of the study which examined the graphic types of the textbooks from Hong Kong, Japan, Mexico, and Canada (1994b) indicated that common graphic types and conventions are found in textbooks across languages. Early, Mohan, and Hooper (1989) also pointed out that using graphics can facilitate the learning of both LI and L2 learners in classrooms which integrated language and content. Acknowledging the full potential of graphic organizers, the current research explores the productive functions of graphic organizers. The use of classification graphics has been integrated into the design of the study. 10 2.6 The knowledge framework and knowledge structure graphics I now turn to discuss Mohan's approach to knowledge structures. Mohan has recently summarized this work in Mohan (in press): To address the issues of education systems which are increasingly multilingual and multicultural, we must look beyond the acquisition of the L2 system and consider education as language socialisation into social practices. This article models social practices as frameworks of knowledge structures which link cultural meanings of the practice to meanings in discourse. The model is situated within systemic functional linguistics and focusses on field (or popularly, 'content') of discourse, showing typical relations between meanings of knowledge structures and language form, the role of atypical or 'metaphorical' relations in constructing advanced knowledge, and how graphics and nonlinguistic media generally can be interpreted as knowledge structures. Educational implications address: integrated approaches to language and content; the connection of language and content standards in education; bridges between learners' languages and cultures; and links to strategies for comprehension in reading and discourse awareness in writing. Current concerns include failures of language assessment to deal adequately with the linguistic construction of content in discourse. Future directions include discourse research strategies to support the potential convergence of multimodal literacy, critical thinking skills, and computer technologies, (p. 1) What, then, is a social practice? A social practice is a unit of culture which involves cultural knowledge and cultural action in a theory/practice, reflection/action relation. Knowledge structures (KSs) are semantic patterns of the discourse, knowledge, actions, artifacts and environment of a social practice (Spradley, 1980, p. 93). In our heuristic model, core KSs of typical social practices include at a general level, classification, principles and values, and at a specific level, description, sequence and choice. (Mohan, in press, p. 2) Mohan places his model within Systemic Functional Linguistics: I situate this model within a systemic functional linguistics approach to language, which views discourse in sociocultural context, learning as a linguistic process, and language learning as language socialisation. SFL is oriented to the description of language as a resource for meaning rather than as a system of rules (see Halliday & Martin, 1993, p. 22-3). Learners are seen as expanding their resources for meaning, their meaning potential, rather than making a transition from errors in rules to the correct form of rules. This has major implications for language assessment as we will see below. Halliday (1994) offers a model of discourse to characterize the relations between text (discourse) and context. A text relates to its context through 11 field, the subject matter of the text or the social practice that is taking place at the time; through tenor, the social relationships that hold among the various participants in the interaction; and through mode, the role of language in the interaction. These three aspects—field, tenor, and mode—are realised in three corresponding types of meaning: ideational, which represents experience, or in other words expresses 'content', the speaker's experience of the world; interpersonal, which enables interaction; and textual, which achieves coherence and connectedness in discourse. The article will focus on field (see Hasan, 1999) though without excluding mode and tenor. (Mohan, in press, p. 2) Systemic functional linguistics is often popularly associated with a genre approach. Mohan contrasts his field approach with a genre approach: There are major differences between field and genre. Field groups texts by topic/social practice; genre groups texts by type. As Martin (Martin, 1997, pp. 12-13) notes, with a field approach, for instance, instructions for an experiment are grouped with doing the experiment, the recount of the experiment and the explanation the experiment illustrates. With a genre perspective, by contrast, the instructions are grouped with similar procedural texts with closely related 'texture' (i.e., a sequence of commands etc.) dealing with any topic. Of course, the science learner relating the explanation of an experiment to the procedure for doing the experiment must not only group texts by field, but must also build scientific understanding by fitting the meaning of these texts together (Mintzes et al., 1998 ). Similarly the researcher in field must explain how a learner can construct a model of a topic or social practice from its texts. How do learners relate explanations of scientific principles to the steps of an experiment? Other field/genre contrasts include field accounts for KSs, while genre accounts for the unfolding of text; anthropologists use the KSs of field for cross-cultural comparison (Werner & Schoepfle, 1987) whereas genres are typically local to a culture. (Mohan, in press, p. 3) 12 Figure 2.2 Form-meaning relations in the field of a social practice CLASSIFICATION PRINCIPLES (e.g. norms, cause-effect, means-end...) EVALUATION - Generic Reference - Being Process - Additive Conjunction - Taxonomy Lexis - Generic Reference - Doing Process - Consequential Conjunction - Consequence Lexis - Generic Reference - Sensing Process - Comparative Conjunction - Evaluation Lexis - Specific Reference - Being Process - Additive Conjunction - Attribution Lexis - Specific Reference - Doing Process - Temporal Conjunction - Sequence Lexis - Specific Reference - Sensing Process - Alternative Conjunction - Alternative/ Choice Lexis DESCRIPTION SEQUENCE CHOICE Mohan contrasts different knowledge structures using features of systemic lexicogrammar (see Figure 1): Figure 1 uses four systems of English: reference, transitivity, conjunction, and lexis (Halliday, 1994; Martin, 1992). We distinguish between the two rows (description vs. classification, sequence vs. principles, and choice vs. evaluation) by the contrast between specific reference and generic reference (e.g., 'the cat', 'the rat' vs. 'cats', 'rats'). We distinguish between the three columns (description and classification vs. sequence and principles vs. choice and evaluation) by the contrast in transitivity of the three main process types of Halliday (1994), processes of being (Is a platypus animal?), doing (Does the rat go?) and sensing (the mental world of consciousness), including processes of liking and wanting (Cats like things, I want my teddy bear). The types of conjunction (e.g., Additive: And is a seal...?, Temporal: when the cat has killed it) are based on Martin (1992, Ch. 4), but separating out the subclass of alternation, either...or (ibid., p. 205). The lexical categories of Figure 1 are inspired by Martin (1992, Ch. 5); for example, taxonomy lexis: is a platypus an animal? Listeners and readers use these categories to understand a text as a cohesive whole. Martin grounds his lexical categories explicitly on field and particularly on his notion of field as sets of activity sequences oriented to some global institutional purpose. This gives rise to the following categories (which he illustrates for the field of tennis): Taxonomy—part—whole relations among game-set match; Configuration—Agent Process Medium structure player-serve-ball; Activity sequence—player serve—opponent return—player volley. Our model of social practice leads to additional categories (again illustrated for tennis): Attribution—numerative-thing structure second serve; Consequence—let— 13 replay; Evaluation—foot-fault; Alternative or Choice—server/receiver. In tennis, umpire or referee judgements reflect the institutional order underlying the social activity; they judge foot-faults and lets, call for replays, and monitor the selection of server and receiver roles. (Mohan, in press, p. 4) Mohan asserts that that his 'knowledge structures' model has implications for the analysis of meanings in multimedia: Each of our group of KSs has widely known graphic conventions for representing it, and these appear regularly in academic textbooks. Sequence has time lines and sequence charts, Classification has trees and Venn diagrams, and so on. To explain these conventions we can extend our account of the linguistic construction of field or content meaning with the concept of 'lines of meaning'. Longacre (1990; 1996) argues that different discourse types have different mainlines of development, or lines of meaning. In narrative discourse the line of meaning is a storyline. Similarly, in procedural discourse, the line of meaning is a line of procedure. Any adequate theory of discourse and literacy has to account for how writers/speakers construct lines of meaning and how readers/listeners interpret them. KSs can be viewed as lines of meaning. (Mohan, in press, p. 5) Mohan describes his analysis of graphics and meaning in some detail: When a reader draws a time sequence line to represent a story (e.g., about hiking, climbing, and then skiing), one can say, drawing on standard work in semantics, that the reader represents the sequence of events in the story in a graphic as a series of nodes of words connected with lines, where typically a node represents an event and a line represents the relation of time sequence: (4) Hike—>—Climb — >—Ski In this way, the semantic relation of time sequence is represented in the story-line of the text and in the drawn line of the graphic. This approach to graphic representation, which applies to all of the KSs, has been widely used in linguistic semantics as a way of representing semantic relations, and its general theory is provided by graph theory in mathematics. In the form of transition networks, it has been used in SFL(e.g., Halliday & Martin, 1993, pp. 158-9, Halliday & Matthiessen, 1999, p. 389). As a way of relating language to socio-semiotic systems such as the visual arts, music and dance and other forms of meaning-making behaviour including charts, maps, diagrams, and so on (see Halliday & Matthiessen 1999, p. 602), it treats graphics as a complementary nonlinguistic resource which realises linguistic meanings, and it differs from other approaches within SFLthat treat socio-semiotic systems on the model of language, as if they had their own grammar and semantics (e.g., Kress & Van Leeuwen, 1996). (Mohan, in press, p. 6) 14 Mohan's approach has educational implications for the analysis of graphic literacy across languages: KSs offer vital bridges between languages and cultures in the globalising educational environment. Tang (e.g., 1997) has examined graphic literacy across languages and shown how certain KS graphics (e.g., the water cycle) occur in school textbooks across several languages and cultures and offer potential bridges to scaffold language development for bilinguals. Mohan and Huang (2002) describe a KS approach to the integration of language, content, and culture in Mandarin. Elementary school anglophone students were enabled to discuss their daily lives in Mandarin, describing and classifying themselves, their classmates and their families, and stating their everyday routines, and comparing their lives to that of a Chinese elementary student. An analysis of student writing illustrates how these KSs appear in the grammar and lexis of Mandarin discourse. For an account of the graphic scaffolding process of this course, see Huang (2003) In an account of academic discourse use in bilingual cooperative learning interactions of Mandarin-English bilingual students, Liang and Mohan (2002) describe how students bilingually correlated their prior knowledge of classifications in their first language and culture with those in their second language and culture. (Mohan, in press, p. 7) Mohan identifies future directions for work guided by his model which provide a context for the work of this thesis: Perhaps the most immediate direction concerns the new multimodal literacies, critical thinking skills and computers. A KS perspective provides a discourse link between these areas to support future convergence. Interest in multimodality, e.g., graphic organizers, has increased enormously with the development of the Internet, as a Web search for the phrase will show, though the quality of information is highly variable... A useful overview of graphic organisers is offered by Hyerle (1996; see also www.thinkingmaps.com). Hyerle emphasises: the constructivist role of learners using graphics to organise and communicate about their knowledge, fundamental patterns of thinking, and a common visual language where different patterns of thinking are consistently represented in corresponding graphic primitives. Hyerle's list of basic graphic elements (Hyerle, 1996, pp. 98, 101) includes most of the KSs above, as do a number of other writers. Thus for researchers, a KS analysis offers a discourse approach which integrates these strands; for language educators, it provides support for the coherent development of the potential of the new media. (Mohan, in press, p. 8) 15 2.7 Summary of the chapter This chapter has identified and reviewed the primary research which relates to the current study. It introduced Halliday's notion of a culture of science language and register, and described how science constructs taxonomies or classifications or technical terms. The development of classification discourse is therefore a critical part of becoming scientifically literate, a point made clear in Unsworth's (2001) work. This chapter has also discussed the development of science literacy across languages, introducing Cummins's work on the interdependency principle (Cummins, 1981), and raising issues about this thread of work. The chapter then continued with a brief review of the use of graphics in education and borrowed from Mohan's current work on knowledge structures and graphics to provide the background to the present study. The next chapter will describe the methodology. 16 CHAPTER 3: RESEARCH METHODOLOGY 3.1 An overview of the research methodology This chapter on the research methodology presents two areas: The research design for the collections of texts (a time series design), and the Systemic Functional Linguistics analysis of discourse. I will discuss them in turn, then restate the research questions which this study aims to address. 3.2 Design of the present study The study is organized as a simple interrupted time series design, with one experimental group and multiple observations before and after two treatments. The design is diagrammed in Figure 3.1. Figure 3.1 The design of the study Ol 02 X 03 04 Y 05 06 The students produce six written classification discourses, three in Chinese and three in English. Written classification discourses in Chinese are represented as 01, 03, and 05, and 02, 04, and 06 are written classification discourses in English. Treatment X is a task where the students read a text about the classification of living organisms in their stronger language, Chinese. Treatment Y is a task where the students listen to a one-hour instructional session on writing classification in Chinese, a session which focused on how different language forms are used to realize the knowledge structure of classification in Chinese. Time series designs have been recognized to have at least two sources of invalidity: history, and the interaction of testing and the treatment, in this case X and Y (Borg & Gall, 1989, p. 663). History is the possibility that forces other than the treatment under investigation came to influence the dependent variable at the same time. While it is not 1 7 possible to rule out history, I have no knowledge of any forces other than the treatment that might have influenced the dependent variable. Interaction of testing and treatment, on the other hand, is a consideration in this study since collecting writing from students means that students are given practice in writing, which may lead to improvement in writing. However, I accept the confounding of testing and treatment, and indeed the confounding of treatment X and treatment Y. Recall that the main emphasis of the study is on the "effects," on the comparison and contrast of student writing at the six stages of the time series, rather than the precise causal relations. 3.2.1 The setting The study was conducted in an educational centre in Richmond, Canada, which offered after-school programs to help ESL students improve their English skills to meet the language demands at school. The majority were native Cantonese speakers while a small population spoke Mandarin as their first language. A l l students were placed into various levels of ESL classes based on their results of an entrance placement English test. 3.2.2 The participants Fourteen students (five female and nine male) participated in this study. However, in line with the central research question to explore how students use their first language (Chinese) and their second language (English) to represent the knowledge structure of classification after both the reading task and the instruction session, the analysis will include data from only twelve of the participants (four female and eight male) who were Cantonese speakers and who had chosen to read a Chinese model text of classification and to receive instruction about writing Chinese classifications in the two treatments in the time series design. Al l twelve subjects had immigrated to Vancouver from Hong Kong during the past three to ten years and were stronger in their first language, (Cantonese) Chinese. A l l were high school students studying grade nine to twelve in a Canadian high school in Richmond or Vancouver. Table 3.1 presents the details of their background: 1 8 Table 3.1: The participants Participant Grade Length of time in Canada S 10 6 years PA 11 10 years M A 11 3 years A N 12 5 years W 9 3 years I 12 6 years J 11 4 years s u 10 3 years JO 10 3 years P 12 5 years M 10 5 years C 9 4 years 3.2.3 The data collection The data were collected from a sequence of classification tasks, which involved a pre-reading classification task, a pre-reading writing task, a reading task, a post-reading discussion task, and two post-reading writing tasks, producing two drafts of writing. The following summarizes the data collection procedures: (1) Pre-task: The participants were asked to describe their understanding of a classification tree. (2) Pre-reading task |Pre-R|: The participants were asked to sort the pictures of living organisms into groups based on their background knowledge and to represent their knowledge in a classification tree. After that, they were asked to write up their knowledge about the classification of living organisms in text form, in both Chinese and English, based on the classification tree they had constructed. (3) Reading task: Participants read a text about the classification of living organisms in their preferred (stronger) language, Chinese.1 The Chinese text is written with reference to the "Diversity of Living Organism" web page of HKIEd HAS Center at www.ied.edu.hk/has/bio/dlo/introd/index.htm and a biology textbook used in a Chinese secondary school in Hong Kong, Pang, K.C. (1998): « 4" * ft =f S $ £ ft * 1 » . ff i t W tfj HI tt<> 19 (4) Discussion task: Participants worked in pairs to re-construct what they had learnt from their reading in a classification tree using English or Chinese or both English and Chinese (mixed-code) at their own choice. (5) Post-reading task |Post-R(l)|: With the aid of the classification tree co-constructed in the discussion task, participants were asked to describe in writing the classification of living organisms in both Chinese and English. (6) Post-instruction task (Post-R(2)|: After a one-hour instruction session on writing classification in Chinese (focusing on how different language forms are used to realize the knowledge structure of classification in Chinese), which was conducted by the researcher in Cantonese Chinese, participants were asked to revise the classification tree, if necessary, and to write a second draft on the classification of living things in both English and Chinese. 3.3 Systemic functional analysis of discourse Over the past decade, teaching language in isolation has been recognized to be inadequate in preparing ESL students for content learning in North America. Thus, there has been a shift away from 'isolating' to 'integrating' language and content instruction (ILC), aiming to support ESL students' academic and cognitive development while they are in the process of acquiring English as a second language. According to Mohan (1986), knowledge (culture) can be conceived in terms of 'activities'2 which consist of a combination of 'knowledge structures'—the semantic relations of discourse. Each activity consists of two aspects: (1) the general, theoretical side consists of the knowledge structures of classification, description, and evaluation, and (2) the specific, practical side consists of the knowledge structures of description, sequence, and choice. Most importantly, "knowledge structures are a bridge between the organisation of content knowledge, graphic semiotics, and expository discourse" (p. 105). This serves as the background for analyzing the data of this study from a systemic functional linguistic (SFL) perspective, aiming to Mohan (1986) defines an activity as a unit of cultural knowledge and behavior, a unit of cultural meaning, constructed and nurtured through discourse. 20 reason grammatically about the language of science in the student-constructed classification discourses produced in a series of classification tasks. SFL describes language as a resource for meaning rather than a system of rules, emphasizing the strong relation between form and meaning in discourse. It concerns how people use language to construe and (re)organize their experience of the world (Halliday & Martin, 1993) and is oriented to "what speakers can mean rather than what they can say" (p. 22). As Christie and Unsworth (2000) stated, the functional analysis of. discourse focuses on "how people use language to make meanings... They do this through their selections from the sets of choices that are available in the language systems" (p. 3). Halliday (1979) also emphasizes the need for building a more flexible model of discourse analysis that addresses the relation of form and function: [Traditional grammar] is formal; rigid; based on the notion of rule; syntactic in focus, and oriented towards the sentence. What is needed is a grammar that is functional; flexible; based on the notion of resource; semantic in focus, and oriented towards the text. (p. 186) Based on this orientation of SFL, this study is concerned with texts—how the grammatical devices for classification are used to realize the classification semantic structure in discourse within science registers. Halliday (1994) introduced a SFL model of discourse to reason the relations between discourse and its context in three aspects: field, tenor, and mode. Field is the subject matter of the discourse or the socially recognized activity that is taking place at the time; tenor refers to the social relationships among the various participants in the interaction; mode analyzes the role of language in the interaction. These three aspects are realized in three corresponding types of meaning: ideational, which represents experience; interpersonal, which enables interaction; and textual, which achieves coherence and connectedness in discourse. The analysis of this study concentrates on identifying the ideational meaning of the texts collected in tasks which required classifying living things on Earth, and looking into how the participants used their first and second languages (LI and L2) to represent the classification knowledge structure. As Halliday (1994) underlined, it is the ideational meaning that is closest to the everyday sense of "meaning in the sense of content" (p. 106). To understand how language is integrated with content to construe 21 scientific classifications—the form-function connections—the classification texts are analyzed into carrier/identified - processes - attributes/identifier to display the basic structure of the classification texts. The semantic relations involved in classification may be manifested in a wide range of linguistic forms. Mohan (in press) contrasts different knowledge structures using features of systemic lexciogrammar based on four systems of English: reference, transitivity, conjunction, and lexis. The knowledge structure of classification can be realized using generic reference, being process, additive conjunction, and taxonomy lexis, while description is formed by specific reference, being process, additive conjunction, and attribution lexis. In addition, Derewianka (1991) identified the typical language patterns found in information reports, which describe what an entire class of things is like. According to Derewianka's analysis, the basic linguistic realizations (the language features) that show up in the classification texts include generalized participants, which state a whole class of things; some action verbs (relational processes), which describes the participants' behavior; many linking verbs (relational processes), which link the carrier to its attributes or identified to the identifier; descriptive language, which characterizes the participants; and other language for defining, classifying, comparing, and contrasting. The analysis of the classification texts is built on the previous work of systemic functional linguists (e.g., Collerson, 1994; Derewianka, 1991, 1998; Halliday, 1999; Halliday & Martin, 1993), which suggested directions for analyzing student writing in science, and in particular, how discourse construes students' classification knowledge in science. This study concentrates on identifying how the following listed common features of classification are used by the students in their classification discourse in Chinese and English at different stages of the study to reflect their scientific knowledge of the classification of living organisms. For example, after the pre-reading task, which was designed to elicit the participants' background knowledge, the participants were asked to read a Chinese text, which is their preferred language, about the scientific classification of living organisms. The analysis examined the following language features: 22 a. Reference (Generic) i) Nouns/noun phrases (including modifiers) that create nodes (classes) in the classification tree ii) Noun phrases that define generalized participants, and iii) Examples given to exemplify generalized participants b. Transitivity i) Linking verbs (relational processes): Linking the subject to its larger group ii) Linking verbs (relational processes): Linking the subject to its characteristics (attributes) iii) Linking verbs (relational processes): Linking the subject to its part/components (possessives) iv) Linking verbs (existential processes) v) Action verbs (material processes): Describing behavior c. Lexis i) Verbs/verb phrases for defining and classifying ii) Noun/noun phrases for defining and classifying d. Conjunctions e. Other devices related to defining and classifying f. Descriptive language (factual and precise attribution/quality): items used to elaborate the feature of each category in the classification 3.4 Summary of the discourse analysis procedures This study employs a Knowledge Structure Analysis based on the Knowledge Framework (Mohan, 1986), a conceptual framework for analyzing how language is used to represent meaning in social activities. With respect to this framework, classification, description, principles, sequence, evaluation, and choice are the six categories of semantic structures governing the representation of experiences in social and academic activities. This study focuses on the knowledge structure of classification, and the procedures are as follows: 1. Analyze the reading text based on its ideational meaning of each clause: Clause = carrier/identified - processes - attributes/identifier. 2. Analyze the reading text based on some common features of classification discourse suggested by Collerson (1994), Derewianka (1991, 1998) and Halliday (1999). 23 3. Analyze the ideational meaning of the participants' classification discourse generated in the pre-reading task, post-reading task (draft 1), and post-instruction task (draft 2) into carrier/identified - processes - attributes/identifier. 4. Identify the features of the classification texts produced by the participants in the pre-reading, post-reading, and post-instruction tasks. 5. Identify the changes which occurred in the participants' classification discourse over the course of the study. 6. Analyze the participants' classification discourses into "meaningful units" —classification and elaboration/explanation—to see if they were able to produce a more elaborated classification discourse in both their first and second language after receiving instruction on the features of classification discourse in their stronger language. 3.5 T h e r e s e a r c h ques t ions It should be noted that this is an exploratory study of the relation between learners' language production of scientific classification discourse in Chinese and in English. The main focus is the nature of bilingual learner classification discourse in both languages, since little is known about this. As stated earlier, this study will be given the form of a time-series design, but the main emphasis is on the "effects," the little-known area, rather than on precise causal relations. That is, the main interest will be in the comparison and contrast of student writing at the six stages of the time series. Primary research question: Given Halliday's view that there are commonalities of science register across Chinese and English with respect to scientific classification, how do students use Chinese and English to represent the knowledge structure of classification in the tasks studied? What are some of the similarities and differences? To help respond to the above research question, the following were asked: • How do the participants represent their initial knowledge of the classification in a visual format? Does the reading of a model text of classification change the students' visual representations? 24 • Does reading a model text of classification of Chinese help learners to write more developed classification in Chinese? • Does reading a model text of classification of Chinese help learners to write more developed classification in English? • Does instruction in the discourse features of classification in Chinese change the students' visual representations? • Does instruction in the discourse features of classification in Chinese help learners to write more developed classification in Chinese? • Does instruction in the discourse features of classification in Chinese help learners to write more developed classification in English? Are there positive and/or negative effects of transfer on students' language production in the tasks? Chapter 4 will present the findings of the study and examine these questions in detail. 25 CHAPTER 4: DATA ANALYSIS 4.1 The primary research question The primary research question asked: Given Halliday's view that there are commonalities of science register across Chinese and English with respect to scientific classification, how do students use Chinese and English to represent the knowledge structure of classification in the tasks studied? What are some of the similarities and differences? In order to respond to this question, several different questions were asked, as stated at the end of Chapter 3. These questions will be addressed in this chapter. Before looking at the participants' data, however, it is useful to examine how the model classification texts were constructed in the two languages and to identify the commonalities and differences in the language features used for classification. Section 4.2 will address this point. Section 4.3 will examine how the participants represented their knowledge of the classification in a visual format over the pre-reading and post-reading tasks. Section 4.4 will investigate whether the reading of a model text of classification in Chinese helps learners write more developed classification in Chinese and English. Section 4.5 will highlight the changes in students' visual representations after the instructional session. Section 4.6 will then discuss whether instruction in the discourse features of classification in Chinese helps learners write more developed texts in Chinese and English. Section 4.7 describes the positive and negative effects of transfer on the participants' language production. Finally, Section 4.8 will sum up the findings to address the primary research question stated above. 4.2 The analysis of the language features in the model English and Chinese classification texts As stated in Chapter 3, a Chinese classification text was presented as a model text for the students to read before attempting a rewrite of their initial texts in both Chinese and English. A comparison of the model English and Chinese classification texts (Appendix 1) illustrates that comparable language features were used to form the classification discourse in both English and Chinese in these models. These features consisted of: 26 a. Reference (Generic) i) Nouns/ noun phrases (including modifiers) that create nodes (classes) in the classification tree ii) Noun phrases that define generalized participants, and iii) Examples given to exemplify generalized participants b. Transitivity i) Linking verbs (relational processes): Linking the subject to its larger group ii) Linking verbs (relational processes): Linking the subject to its characteristics (attributes) iii) Linking verbs (relational process): Linking the subject to its part/components (possessives) iv) Linking verbs (existential processes) v) Action verbs (material processes): Describing behavior c. Lexis i) Verbs/verb phrases for defining and classifying ii) Nouns/noun phrases for defining and classifying d. Conjunctions e. Other devices related to defining and classifying f. Descriptive language (factual and precise attribution/quality): items used to elaborate the feature of each category in the classification These language features will be used to illustrate the development of the participants' classification texts and to address the research questions. 4.3 Student-constructed classification trees in the pre-reading and the post-reading tasks Believing that graphic organizers might help students organize and represent their knowledge about the classification of living things, the researcher asked participants to construct a classification tree either their L I , their L2, or both LI and L2 (bilingual) before each writing task. Table 4.1 summarizes the students' language choice in the pre-reading and post-reading classification trees. Graphic I - Self-constructed by the participants in the pre-reading task Graphic 2 - Co-constructed by the participants in pairs in the post-reading task 27 Table 4.1. Language(s) used in the student-constructed pre-reading and post-reading classification trees. Graphic 1 Graphic 2 A+S - Bilingual S - English PA - Bilingual PA+MA - Bilingual M A - Bilingual A N - Chinese AN+W - Chinese W - English 1 - Chinese I+J - Chinese J - Bilingual SU - Chinese SU+JO - Chinese JO - Bilingual P - English P+M - Bilingual M - English C - English C+L - English 1 Graphic 1 Graphic 2 English 5 0 Chinese 3 3 Bilingual 4 2 How did the participants represent their initial knowledge of the classification in a visual format? Did the reading of a model text in Chinese change the students' visual representation? The classifications shown in Figure 4.1 represented the participants' "common sense" classification, while the groupings shown in Figure 4.2 and Figure 4.3 reflected the understanding of the "scientific" classification of living organisms which the participants had learned from reading the scientific text. Such informal (common sense) and formal (scientific) classifications displayed in these classification trees were realized in both the Chinese and the English written discourse in the pre-reading and post-reading writing task respectively. 28 Figure 4.1 Example of a student-constructed classification tree in the pre-reading task (Graphic drawn by Participant I) Figure 4.2. Example of a student-constructed classification tree in the post-reading task (Graphic co-constructed by Participant I and Participant J) 29 The result showed that both the LI and the L2 had been used by the participants as language resources to help represent their knowledge of the world visually on the classification trees, which in this case were their "common sense classification of organisms" and the newly acquired "scientific classification of living organisms." Figure 4.3. Example of a student-constructed classification tree in the post-reading task (Graphic co-constructed by Participant PA & M A in both Chinese and English in the post-reading task ^ ft* • ^ 3 i » i ^ ® * * * * * 0 Cf|«cV»v»~0 <3 ^ f t ^ ^ ^ M O 0 4.4 Examining the effects of reading a model classification text Does reading a model text of classification of Chinese help learners to write more developed classification in Chinese? Does it help learners to write more developed classification in English? This section will look at these questions quantitatively and qualitatively. 30 4.4.1 A comparison of the pre-reading and post-reading classification discourse Based on their self-constructed classification trees in the pre-reading task and the co-constructed classification trees drawn in pairs after the reading the task, the participants produced four written texts, two in Chinese and two in English, to represent their knowledge of the classification of living things in the pre-reading and post-reading task. The following sections compare the findings of these pre-reading and the post-reading texts. 4.4.1.1 Quantitative results: Descriptive No significant result was found in the comparison. Eight of the participants showed a slight increase in the total number of language features used for classification in their Chinese discourse, while four displayed a decrease. Seven of the participants showed a slight increase in the total number of language features in their English classification discourse, while five students demonstrated a drop in the total number of language features used in their English classification discourse. Table 4.2 summarizes the number of language features found in the pre-reading and post-reading Chinese classification discourse task, and Table 4.3 offers the full details of features used. Similarly, Table 4.4 summarizes the number of language features found in the pre-reading and post-reading English classification task, and Table 4.5 presents the full details. Table 4.2 A summary of the total number of language features found in the pre-reading and post-reading Chinese classification discourse Participant Pre-reading Classification Discourse Post-reading Classification Discourse Result S 53 86 + PA 164 93 _ MA 131 120 AN 63 142 + W 62 102 + I 102 142 + J 83 111 + SU 67 63 _ JO 63 84 + P 65 38 M 13 45 + C 48 63 + 31 03 ro Participants' Chinese Classification Discourse Pre-R Post-R Pre-R Post-R Pre-R Post-R Pre-R Post-R Pre-R Post-R Pre-R Post-R Pre-R Post-R Language Features S S PA PA MA MA AN A N W W 1 I J J I Reference (Generic) A. Nouns/ Noun Phrases (nodes) 17 37 50 28 32 36 15 38 23 26 25 44 28 42 B. Noun Phrases (definition) 0 0 7 0 0 0 0 0 0 0 0 0 0 0 C. Examples 1 12 23 19 20 14 14 37 18 36 18 17 23 16 18 49 80 47 52 50 29 75 41 62 43 61 51 58 II. Transitivity A. Linking Verbs [Relational Process]: (identified - identifier) 1 5 13 6 6 13 1 2 0 0 8 19 0 10 B. Linking Verbs [Relational Process] (carrier-attributes) 3 0 1 0 8 2 0 0 0 0 1 0 0 0 C. Linking Verbs [Relational Process]: (possessives) 7 6 11 2 7 10 0 4 6 6 4 10 12 18 D. Linking Verbs [Existential Process] 0 0 0 0 2 0 0 0 0 0 0 0 0 0 E. Action Verbs [Material Process]: Describing behavior 9 0 11 0 7 0 7 14 1 0 4 1 1 0 20 11 36 8 30 25 8 20 7 6 17 30 13 28 III. Lexis A. Verbs/Verb Phrases for defining and classifying 3 3 5 5 2 3 1 5 2 3 3 2 2 4 B. Nouns/Noun Phrases for defining and classifying 3 18 22 19 20 22 10 25 2 13 19 34 9 12 6 21 27 24 22 25 11 30 4 16 22 36 11 16 IV. Conjunctions -A. Additive 6 4 8 4 9 9 8 17 7 4 12 12 7 9 B. Alternatives 0 0 1 0 0 0 0 0 0 0 0 0 0 0 C. Contrast 0 0 0 0 2 2 0 0 0 0 0 0 0 0 D. Others 0 0 0 0 1 0 0 0 0 0 0 0 0 0 6 4 9 4 12 11 8 17 7 4 12 12 7 9 V. Other Devices related to Defining and Classifying 1 1 12 10 9 4 7 0 3 14 8 3 1 0 1 1 12 10 9 4 7 0 3 14 8 3 1 0 VI. Descriptive Language (attribution/quality) 2 0 0 0 6 5 0 0 0 0 0 0 0 0 2 0 0 0 6 5 0 0 0 0 0 0 0 0 Total: 53 86 164 93 131 120 63 142 62 102 102 142 83 111 Pre-R = Pre-reading Task Post-R = Post-reading Task H 6= cr a U> o , 70 u> a> co co CO c_ CO ca O (-»• -+> ion i f sr a> to co 3 7? EL CO CO O <—f =r -o S5 5. •5' to 3 r f CO -a a> a S» Q . 3' Cro P 3 CL ~o o CO r t -n n> r» a . 3' era O 3^ 5' a> CO Participants' Chinese Classification Discourse Pre-R Post-R Pre-R Post-R Pre-R Post-R Pre-R Post-R Pre-R Post-R Language Features SU SU JO JO P P M M C C 1 Reference (Generic) A. Nouns/ Noun Phrases (nodes) 26 32 16 32 13 13 7 23 19 18 B. Noun Phrases (definition) 0 0 1 1 0 0 0 0 0 0 C. Examples 3 0 17 10 9 1 0 7 8 10 29 32 34 43 22 14 7 30 27 28 II. Transitivity A. Linking Verbs [Relational Process]: (identified - identifier) 2 5 0 6 0 2 0 0 0 3 B. Linking Verbs [Relational Process] (carrier-attributes) 0 0. 3 1 0 0 0 0 0 1 C. Linking Verbs [Relational Process]: (possessives) 3 3 2 9 5 6 1 4 4 15 D. Linking Verbs [Existential Process] 1 1 0 0 0 0 0 0 0 0 E. Action Verbs [Material Process]: Describing behavior 0 0 • 10 0 9 0 2 0 7 6 6 9 15 16 14 8 3 4 11 25 III. Lexis A. Verbs/Verb Phrases for defining and classifying 9 4 1 2 3 2 1 3 2 2 B. Nouns/Noun Phrases for defining and classifying 8 9 2 9 17 6 1 1 4 4 17 13 3 11 20 8 2 4 6 6 IV. Conjunctions A. Additive 15 8 4 8 4 7 1 6 4 3 B. Alternatives 0 0 2 1 0 0 0 0 0 1 C. Contrast 0 0 1 0 0 0 0 0 0 0 D. Others 0 0 2 0 0 0 0 0 0 0 15 8 9 9 4 7 1 6 4 4 V. Other Devices related to Defining and Classifying 0 1 1 3 5 1 0 1 0 0 0 1 1 3 5 1 0 1 0 0 VI. Descriptive Language (attribution/quality) 0 0 1 2 0 0 0 0 0 0 0 0 1 2 0 0 0 0 0 0 Total: 67 63 63 84 65 38 13 45 48 63 Pre-R = Pre-reading Task Post-R = Post-reading Task Table 4.4 A summary of the total number of language features found in the pre-reading and post-reading English classification discourse Participant Pre-reading Classification Discourse Post-reading Classification Discourse Result S 49 73 + PA 166 63 _ MA 126 91 _ AN 48 131 + W 72 100 + • I 121 139 + J 120 124 + SU 53 60 + JO 59 57 P 105 35 _ M 12 31 + C 53 48 -34 cr Participants' English Classification Discourse Pre-R Post-R Pre-R Post-R Pre-R Post-R Pre-R Post-R Pre-R Post-R Pre-R Post-R Pre-R Post-R Language Features S S PA PA MA MA AN AN W W 1 1 J J I Reference (Generic) A. Nouns/ Noun Phrase (Nodes) 21 35 55 29 36 40 15 42 25 26 38 42 33 46 B. Noun Phrases (definition) 0 0 11 2 I 1 0 0 0 0 0 0 0 0 C. Examples 2 8 23 14 21 6 14 41 14 43 20 16 23 17 23 43 89 45 58 47 29 83 39 69 58 58 56 63 II. Transitivity .A. Linking Verbs [Relational Process]: (identified - identifier) 2 6 16 1 4 7 1 2 6 1 6 23 9 16 B. Linking Verbs [Relational Process] (carrier-attributes) 1 0 1 0 4 1 0 0 0 0 1 0 0 0 C. Linking Verbs [Relational Process]: (possessives) 0 1 7 1 1 4 7 0 6 1 4 1 1 4 D. Linking Verbs [Existential Process] 4 1 1 0 2 2 0 3 0 1 0 0 1 0 E. Action Verbs [Material Process]: Describing behavior 1 0 15 0 12 1 0 15 1 1 13 0 5 0 8 8 40 2 23 15 8 20 13 4 24 24 16 20 III. Lexis A. Verbs/Verb Phrases for defining and classifying 2 3 6 4 8 3 1 4 1 4 5 5 5 4 B. Nouns/Noun Phrases for defining and classifying 5 11 6 5 8 7 2 5 6 5 8 39 18 28 7 14 12 9 16 10 3 9 7 9 13 44 23 32 IV. Conjunctions A. Additive 3 7 10 4 13 9 8 19 13 4 11 9 15 7 B. Alternatives 1 0 3 0 4 0 0 0 0 0 0 0 0 0 C. Contrast 0 0 0 0 2 0 0 0 0 0 0 0 0 0 D. Others 0 0 0 0 0 1 0 0 0 0 0 1 0 0 4 7 13 4 19 10 8 19 13 4 11 10 15 7 V. Other Devices related to Defining and Classifying 5 1 10 3 6 1 0 0 0 14 14 3 10 2 5 1 10 3 6 1 0 0 0 14 14 3 10 2 VI. Descriptive Language (attribution/quality) 2 0 2 0 4 8 0 0 0 0 1 0 0 0 2 0 2 0 4 8 0 0 0 0 1 0 0 0 Total: 49 73 166 63 126 91 48 131 72 100 121 139 120 124 2. 70 as n co CO =1 CO Sss O - 1 n> Q_ fa to' 3 O 05 g CO —I S2 60 ° O Pre-R = Pre-reading Task Post-R = Post-reading Task 3 cro m 3 CTQ co Participants' English Classification Discourse Pre-R Post-R Pre-R Post-R Pre-R Post-R Pre-R Post-R Pre-R PostR Language Features SU SU JO JO P P M M C C 1 Reference (Generic) A. Nouns/ Noun Phrase (Nodes) 23 34 21 34 23 14 7 20 20 18 B. Noun Phrases (definition) 0 0 1 0 0 0 0 0 0 1 C. Examples 5 3 6 5 19 3 0 0 11 6 28 37 28 39 42 17 7 20 31 25 11. Transitivity A. Linking Verbs [Relational Processl: (identified - identifier) 4 2 5 3 6 4 0 3 2 3 B. Linking Verbs [Relational Processl (carrier-attributes) 0 0 1 0 0 0 0 0 0 0 C. Linking Verbs [Relational Processl: (possessives) 0 1 1 1 3 1 0 1 1 2 D. Linking Verbs [Existential Processl 0 1 0 0 1 0 0 0 1 0 E. Action Verbs [Material Processl: Describing behavior 1 0 8 1 15 0 2 0 7 7 5 4 15 5 25 5 2 4 11 12 111. Lexis A. Verbs/Verb Phrases for defining and classifying 5 4 2 3 2 1 1 2 2 2 B. Nouns/Noun Phrases for defining and classifying 5 3 1 2 19 7 1 1 6 0 10 7 3 5 21 8 2 3 8 2 IV. Conjunctions A. Additive 10 9 10 5 11 2 1 3 3 5 B. Alternatives 0 0 0 0 0 1 0 0 0 2 C. Contrast 0 0 0 0 3 1 0 0 0 0 D. Others 0 0 2 0 0 0 0 0 0 0 10 9 12 5 14 4 1 3 3 7 V. Other Devices related to Defining and Classifying 0 2 0 3 3 1 0 1 0 1 0 2 0 3 3 1 0 1 0 1 VI. Descriptive Language (attribution/quality) 0 1 1 0 0 0 0 0 0 1 0 1 1 0 0 0 0 0 0 1 Total: 53 60 59 57 105 35 12 31 53 48 cr CD in 'o o 3 C CD a. Pre-R = Pre-reading Task Post-R = Post-reading Task 4.4.1.2 Quantitative results: Statistical The pre-reading Chinese was compared with the post-reading Chinese using the Wilcoxon Signed Ranks Test on SPSS. The difference was not statistically significant. The pre-reading English was compared with the post-reading English using the same statistical test. As with the Chinese results, the difference was not statistically significant. Table 4.6 The test statistics: Comparison of pre-reading and post-reading Chinese and English classification texts POSTC - PREC POSTE - PREE z -1.570(a) -.314(a) Asymp. Sig. (1-tailed) .058 .377 a Based on negative ranks b Wilcoxon Signed Ranks Test PREC = Pre-reading Chinese POSTC = Post-reading Chinese PREE = Pre-reading English POSTE = Post-reading English 4.4.1.3 Qualitative results A close look at the data—the classification texts produced by the participants in the pre-reading and the post-reading task—from a qualitative perspective revealed three key points about the participants' performance in the pre-reading and the post-reading tasks. The following examples in Table 4.7, four texts from one of the participants, SU, will be given to illustrate these points. 37 Table 4.7 Examples of SU's texts Text 1 - Participant SU's Pre-reading Chinese Classification Discourse Carrier/Identified m Basically, things on Earth non-living things 5 Stone For man-made Living things Processes existential/relational/material can be classified into respectively are can be further divided into It* is have Plants Similarly, animals can also be divided into respectively are Attributes/Identifier two big categories. living things and non-living things. natural and man-made. natural non-living things. tools, appliances, etc. three big types. can be classified as t H W # f t can also be divided into plant, animal and bacteria. flowering plants and non-flowering plants. two big kinds. "Ell They mammals and non-mammals. non-mammals can be further divided into insects, birds, marine living things and poultry. bacteria also exist classify beneficial and not beneficial. Textual Carrier/Identified Processes existential/relational/material Attributes/Identifier Basicly [Basically], things on earth can be classified in two groupfsl. They [The two groups! are living and non-living thing. Non-living thing can be classified in nature and non-nature. Stone is a good example for nature, and non-nature provided tools and electricity, etc. Living things can be classified in plant, animal, and jerk [germl. Mammal and non-mammal are two big groups for animal, and mammal can be classifield [classified! detailly in bugs, birds and sea living thing. Plant can also be classified in non-flower and flower. Grass and trees are examples for non-flower types. 38 Table 4.7 (continued) Text 3 - Participant SU's Post-reading Chinese Classification Discourse Textual Carrier/Identified Processes existential/relational/material Attributes/Identifier Things on Earth can be classified into mm respectively is two main types, living things and non-living things. rffi while living things even can be classified as five main types. tSffi they [The Jive main types] ft are bacteria, protist, fungus, plant and animal. inside the animal kingdom, have two main types, vertebrates and invertebrates. vertebrates mainly have & , A , mn, mmmm. fish, bird, reptile, amphibians and myriapods. m while ?&m, un, tfcfi coelenterates, jlatworms, roundworms, molluscs, echinoderms, ringed worms, and arthropods ffift then is invertebrates 's kinds. arthropods even is occupying m&wiyo&j 80%, all animals' 80% // [Arhtropods] can be further classified in detail as myriapods, crustaceans, myriapods, and insects. Text 4 -f Participant SU's Post-reading English Classification Discourse Textual Carrier/Identified Processes existential/relational/material Attributes/Identifier Things on earch [Earth] can be classifield [classified] in living thing and non-living thing. Living thing can be classified in bacteria kingdom, protist kingdom, fungus kingdom, plant kingdom, and animal kingdom. In animal kingdom, there are two big groups. and they [The two groups] are vertebrates and invertebrates. Vertebrates have reptiles, mammal, amphian [amphibian], fish and bird. Invertebrates cassifiled [classified] in ringed worms, molluscs, echinoderms, anthopods, crustaceans, coelenterates, flatworms, and round worms. Anthropods is the biggest group in animal kingdom, and it [arthopods] can be classifield [classified] in insects, arachinds [arachnids], myriapods, and crustaceans; for instance, lobster, scorpion, and cetepede, etc. 39 4.4.1.3.1 Common sense versus scientific classification Most participants classified living organisms based on their "common sense" in the pre-reading task, demonstrating that they had possessed limited understanding about the scientific classification of living things before the reading task. However, their post-reading classification discourse reflected the "scientific" rather than the "common sense" classification of living organisms, showing that the participants had learned the "scientific" classification of the living organisms from the reading text. Participant SU, in a manner similar to the other participants, classified all living things into three main types: plants, animals, and jerk [germs]; animals were further grouped under two main categories: mammals and non-mammals. However, after reading about the scientific classification of living things, SU, like the other participants, was able to state in the post-reading classification discourses the formal classifications involved in the topic. 4.4.1.3.2 Transfer of the content structure across the L I and L 2 In both the pre-reading and post-reading task, the participants were required to express their knowledge about the classification of living organisms in a classification tree and in two written texts, one in their LI and the other in their L2. The results showed that the classification tree and the LI and L2 classification discourse produced by the participants consistently reflected their understanding about the classification of living organisms at different stages in the study. The participants' "common sense" classification was shown in i) the LI pre-reading classification discourse, ii) the L2 pre-reading classification discourse, and iii) the self-constructed classification tree in the pre-reading task, while the "scientific" classification was displayed in i) the LI post-reading classification discourse, ii) the post-reading L2 classification discourse, and iii) the co-constructed classification tree in the post-reading task, despite the fact that the learning about the scientific classification of living organisms was gained from only their first language, Chinese. This demonstrated that the participants were able to transfer their knowledge about the "scientific" classification of living organism from their LI to their L2 and to express the content structure in both Chinese and English. Given some help with the vocabulary, the participants were able to reproduce 40 the scientific classification of living organisms learned in Chinese not only in Chinese but also in English with the use of a classification tree. As reflected in all participants' texts, SU's pre-reading Chinese and English classification discourse presented his common-sense classification, while both of his post-reading Chinese and English classification discourse presented the formal classification of living organisms, despite having learned about the scientific classification of living organisms from only his L I , Chinese. 4.4.1.3.2 LI versus L2 classification discourse In the pre-reading task, most participants demonstrated that they had acquired the fundamental language skills required to represent the knowledge structure of classification in both their LI (Chinese) and their L2 (English). The following tables present examples of the language features of classification found in one participant's pre-reading classification discourse. Examples from SU's classification discourse will be given, and where no examples can be found in his discourse, examples from other participants will be provided. Table 4.8, Examples from a participant's pre-reading Chinese classification discourse a. Reference (Generic) i) Nouns/ noun phrases (including modifiers) that create nodes (classes) in the classification tree Student SU Textual Carrier/Identified Processes existential/relational/material Attributes/Identifier Mm Basically, things on Earth can be classified into respectively are two big categories. living things and non-living things. (Pre-reading, Chin, 9.1) %M Living things can also be divided into respectively are three big types. plant, animal and bacteria. (Pre-reading, Chin, 9.5) B5 Also non-mammals can be further divided into insects, birds, marine living things and poultry. (Pre-reading, Chin, 9.9) 41 Table 4.8 (continued) ii) Noun phrases that defines generalized participants Student PA: Textual" Carrier/Identified- Processes • existential/relaUonal/matenal Attributes/Identifier :V-P Flowering plants is [ [ # H « « | ] ] W » . can blossom's plants. (Pre-reading, Chin, 3.3) omnivorous insects SfcJS: are mm [ [#^&mx#i£Kft ! iJ i f i i f3&f-'-'jMM. existenual/telational/matenal': Attributes/Identifier For man-made have tools, appliances, etc. (Pre-reading, Chin, 9.4) 42 Table 4.8 (continued) iv) Linking verbs [existential processes] Participant MA: :" Textual Carrier/Identified Processes ' existentwIAelational/matenal "Attributes/Identifier The whole ecosystem exists six big kinds of living things: marine living things, terrestrial living things, birds insect type, plant type andfungus type (Pre-reading, Chin, 4.5) v) Action verbs [material processes]: Describing behaviour Student M A: Textual Carrier/Identified Processes existenluilittlaiiorial/naterial Attributes/identifier They [Birds] m can live. on land and in ocean (Pre-reading, Chin, 4.13) c. Lexis i) Verbs/ verb phrases for defining and classifying Student SU: Textual^ Carrier/Identified Processes , exist^tul^eUtionalAiiatenal Attributes/Identifier Living things can also be divided into respectively are three big types. plant, animal and bacteria. (Pre-reading, Chin, 9.5) mmm, mm Similarly, animals tn*r## can also be divided into two big kinds. (Pre-reading, Chin, 9.7) ii) Noun/ noun phrases for defining and classifying Student SU: Textual Carrier/Identified Processes '. exiktentUl/telational/rmteriat Attributes/Identifier - ; Living things can also be divided into respectively are three big types. plant, animal and bacteria. (Pre-reading, Chin, 9.5) 43 Table 4.8 (continued) d. Conjunctions Student SU: Textual Carrier/Identified Processes existaitml/relalionalAnaterial, Attributesfldentifier m Also non-mammals can be further divided into a&, fc, insects, birds, marine living things and poultry. (Pre-reading, Chin, 9.9) fttft, Finally, bacteria also exist classify beneficial and not beneficial. (Pre-reading, Chin, 9.10) e. Other devices related to defining and classifying Student MA: Textual Carrier/Identified Processes <; .existentialA l^ational/rhat^ al:; Attributes/Identifier do not have life's things [Non-living things] like telephone, coins, etc. (Pre-reading, Chin, 4.2) Birds-type such as ducks are amphibians. (Pre-reading, Chin, 4.12) f. Descriptive language (factual and precise attribution/ quality) - items used to elaborate the feature of each category in the classification Student MA: Textual Carrier/Identified Processes emstenUal/relauonal/material Attributes/Identifier Some reptile type of insects ft is have poison [poisonous], such as scorpion, centipedes, etc. (Pre-reading, Chin, 4.17) 44 Table 4.9 Examples from a participant's pre-reading English classification discourse a. Reference (Generic) i) Nouns/ noun phrases (including modifiers) that create nodes (classes) in the classification tree Student SU: Textual Carrier/Identified Processes existential/relational/mate rial Attributes/Identifier Mammal and non-mammal are two big groups for animal, (Eng, Pre-reading, 9.7) and mammal can be classifield [classified] detailly in bugs, birds and sea living thing. (Eng, Pre-reading, 9.8) ii) Noun phrases (including modifiers) that define generalized participants Student PA: Carrier/Identified Processes • existential/relational/mater ial Attributes/Identifier Mammals are those [[that behave much like humans]], such as lions or mouse; (Eng, Pre-reading, 3.7) microorganisms are animals [[which are too small to be noticed with human eye]], such as planarian; (Eng, Pre-reading, 3.11) amphibians are animals [[that live underwater or on the land]], such as frogs: (Eng, Pre-reading, 3.12) iii) Examples given to exemplify generalized participants Student SU: Carrier/Identified Processes existentialA-cIational/material Attributes/Identifier Stone is a good example for nature, Grass and trees are examples for non-flower types. 45 Table 4.9 (continued) b. Transitivity i) Linking verbs [relational processes]: Linking the subject to its larger group Carrier/Identified Processes existeritial/relatiohal/mater ial Attributes/Identifier They f The two groups] are living and non-living thing. (Eng, Pre-reading, 9.2) Stone is a good example for nature, (Eng, Pre-reading, 9.4) Mammal and non-mammal are two big groups for animal, (Eng, Pre-reading, 9.7) Grass and trees are examples for non-flower types. (Eng, Pre-reading, 9.10) ii) Linking verbs [relational processes]: Linking the subject to its characteristics (attributes) Student fvl rA-Textual Carrier/Identified Processes existential/relational/mate' rial Attributes/Identifier and One kind of insects such as scorpion and planarian are usually slug poisonous on the ground. (Eng. Pre-reading, 4.16) snails are always slug on ground; (Eng, Pre-reading, 4.17) however it is non-posionous. (Eng, Pre-reading, 4.18) iii) Linking verbs [relational processes]: Linking the subject to its part/components (possessives) Student AN: Carrier/Identified Processes existential/relational/mater ial Attributes/Identifier Living things have lions and penguin. (Eng, Pre-reading, 5.8) Insects have ants and butterfly. (Eng, Pre-reading, 5.9) Student L: Carrier/Identified Processes existential/relational/mater ial-Attributes/Identifier The mammals living on land consist of lions and kangaroo; (Eng, Pre-reading, 14.9) in ocean include sharks and whales. (Eng, Pre-reading, 14.10) Some amphibians contain with shells like crab. (Eng, Pre-reading, 14.13) iv) Linking verbs [existential process] Student MA: Carrier/Identified Processes existeirtUlA-elanonalAnfllenal Attributes/Identifier There are mainly two major groups of things on earth - living and non-living things. (Eng. Pre-reading, 4.1) 46 Table 4.9 (continued) Student J: Carrier/Identified Processes exiStential/reloUonal/matenol Attributes/Identifier In the fish section, there are sharks and seahorse. (Eng, Pre-reading, 8.8) v) Action verbs [material processes]: Describing behavior Student I: Textual Carrier/Identified Processes existential/relational/material Attributes/Identifier In microorganism, some live in water, like leech (Eng, Pre-reading, 7.6) and some live in other animals' body. (Eng, Pre-reading, 7.7) In insects, some of them will fly, for example, bees and butterfly. (Eng, Pre-reading, 7.9) and some live in a group. (Eng, Pre-reading, 7.11) and For birds, owls are sleep very active at night at morning. (Eng, Pre-reading, 7.13) Not all the birds can fly. (Eng, Pre-reading, 7.14) Ostrich is can't fly. the one that (Eng, Pre-reading, 7.15) c. Lexis i) Verbs/ verb phrases for defining and classifying Student SU: Textual Carrier/Identified "•Processes,. lexistcirtiaiWateial/nateruU Attributes/Identifier Basicly [Basically], things on earth can be classified in [to] two grouprs]. (Eng, Pre-reading, 9.1) Living things can be classified in [to] plant, animal, and jerk [germ]. (Eng, Pre-reading, 9.6) and mammal can be classifield [classified] detailly in[to] bugs, birds and sea living thing. (Eng, Pre-reading, 9.8) ii) Noun/ noun phrases for defining and classifying Student SU: Carrier/Identified Processes; existential/relationalAnaterial Attributes/Identifier Basicly [Basically], things on earth can be classified in two group[s]. (Eng, Pre-reading,9.1) Mammal and non-mammal are two big groups for animal, (Eng, Pre-reading, 9.7) Grass and trees are examples for non-flower types. (Eng, Pre-reading, 9.10) 47 Table 4.9 (continued) d. Conjunctions Student S U: Textual Carrier/Identified Processes •;e t^entia I/relitional/material Attributes/Identifier They [The two groups] are living and non-living thing. (Eng, Pre-reading, 9.2) Mammal and non-mammal are two big groups for animal, (Eng, Pre-reading, 9.7) and mammal can be classifield [classified] detailly in bugs, birds and sea living thing. (Eng, Pre-reading, 9.8) e. Other devices related to defining and classifying Student I: like Textual Carrier/Identified Processes existential/relational/mat erial Attributes/Identifier In microorganism, some rmicro]org;anisms live in water, like leech (Eng, Pre-reading, 7.6) Some contain poison, like scorpion (Eng, Pre-reading, 7.10) for example Carrier/Identified Processes existential/relational/mater-ial Attributes/Identifier In amphibian, we can find animals that have shell - for example, crab. (Eng, Pre-reading, 7.3) For example, hookworm IS the one. (Eng. Pre-reading. 7.8) f. Descriptive language (factual and precise attribution/ quality) - items used to elaborate the feature of each category in the classification Student PA: Carrier/Identified Processes existetrtial/relational/mater ial .Attributes/Identifier reptiles are those cold-blooded; such as snakes; (Eng, Pre-reading, 3.8) Student I: Textual Carrier/Identified Processes Attributes/Identifier existetrtial/relational/mat erial and For birds, owls are very active at night sleep at morning. 48 Similar features of the classification knowledge structure found in the pre-reading task were used to form the post-reading classification discourse in their L I , Chinese, and their L2, English. This further demonstrated that most participants were able to select basic devices from both their first and second languages to represent the semantic knowledge of classification. However, although the participants were able to express the knowledge structure of classification using some of the key features of classification, the analysis showed that the participants' classification in both the pre-reading and the post-reading tasks had not been elaborated by details: a. The classes were not defined; not many noun phrases were used to define generalized participants. b. The categories in the classification were not elaborated: • Not many linking verbs were used to link the generalized participants to its characteristics (attributes). • Limited descriptive language was used to describe the features (the factual and precise attribution/quality) of each category in the classification. • Not many action verbs (material process) were used to describe the subject's behavior. The comparison showed that the participants were able to classify the living organisms into formal categories in the post-reading task, offering evidence for their progression from classifying the organisms into categories of their common sense to scientific classifications. Nevertheless, they lacked knowledge in producing a more illustrated classification discourse constructed with an expanded choice of devices used for classification. 4.5 Analysis of the post-instruction classification trees Did instruction in the discourse features of classification in Chinese change the students' visual representations? The classification trees co-constructed by the participants in pairs earlier in the task sequences were returned to the participants in the post-instruction task as a reference. They were allowed, though not required, to add details to the classification tree. Six participants completed the task and all added more elaborated 49 information of classification to the post-reading classification tree they had drawn. This illustrated how the students were further elaborating the details of their classification in their written work, showing more complete information of classification visually in the classification trees. Figure 4.4 is an example of this more elaborated classification tree. Figure 4.4 An example of the classification tree modified by Participant M A in the post-instruction task ***** ^ CD ^*\\K\bl Ut{\i['z**{\ V***. 4.6 Examining the effects of instruction Did instruction in the discourse features of classification in Chinese help learners write more developed classification in Chinese? Did it help learners write more developed classification in English? This section will examine both quantitatively and qualitatively, the effects of the instruction—given in the participants'first language—on their subsequent Chinese and English classification texts. 50 4.6.1 A comparison of the post-reading and post-instruction classification discourse The change noted in the comparison of the pre-reading and post-reading classification discourse illustrated that it is likely that content knowledge can be transferred from the learners' first language, Chinese, to their second language, English. However, the detailed analysis of both the Chinese and English pre-reading and post-reading discourse indicated that the participants in this study needed instruction to write a more illustrated classification discourse. Therefore, a follow-up instructional session was held to teach the participants the language features of classification, focusing on the various linguistic choices that are available in their first language, Chinese, for improving their writing of classification discourse. After the instructional session, the participants were asked to revise their classification tree, if they thought it was necessary (see Section 4.5), and to revise the draft they had produced in the post-reading task. The next sections will present the findings of the comparison of the post-reading and post-instruction classification discourse. 4.6.1.1 Quantitative results: Descriptive The comparison of the Chinese post-reading and the Chinese post-instruction classification discourse showed a large increase in the total number of language features used in the participants' post-instruction classification discourse written in their L I , Chinese. These results are summarized in Table 4.10 and described in detail in Table 4.11. The comparison of the English post-reading and the English post-instruction classification discourse also showed a measurable increase in the total number of language features used by the participants in their L2 post-instruction classification discourse, despite the fact that instructions about how to write classification were given only in their LI and not in their L2. Table 4.12 summarizes these findings, and Table 4.13 offers a detailed description of the language features. 51 Table 4.10 A summary of the total number of language features found in the post-reading and post-instruction Chinese classification discourse Participant Post-reading Classification Discourse Post-instruction Classification Discourse Result P A 93 349 + M A 120 329 + A N 142 295 + J 111 145 + SU 63 201 + JO 84 154 + 52 s 3 CT a 4^ CO Participants' Chinese Classification Discourse Post-R Post R(2) Post-R Post R(2) Post-R Post R(2) Post-R Post R(2) Post-R Post R(2) Post-R Post R(2) Language Features PA PA M A MA AN AN J J SU SU JO JO I Reference (Generic) A. Nouns/ Noun Phrases (nodes) 28 84 36 85 38 75 42 33 32 60 32 53 B. Noun Phrases (definition) 0 4 0 6 0 5 0 1 0 3 1 1 C. Examples 19 32 14 34 37 39 16 14 0 14 10 16 47 120 50 125 75 119 58 48 32 77 43 70 II. Transitivity A. Linking Verbs [Relational Process]: (identified - identifier) 6 14 13 16 2 13 10 5 5 13 6 7 B. Linking Verbs [Relational Process] (carrier-attributes) 0 14 2 4 0 9 0 3 0 10 1 9 C. Linking Verbs [Relational Process]: (possessives) 2 30 10 19 4 29 18 22 3 21 9 10 D. Linking Verbs [Existential Processl 0 0 0 0 0 0 0 0 1 0 0 0 E. Action Verbs [Material Processl: Describing behavior 0 30 0 29 14 26 0 13 0 6 0 9 8 88 25 68 20 77 28 43 9 50 16 35 III. Lexis A. Verbs/Verb Phrases for defining and classifying 5 9 3 7 5 10 4 4 4 5 2 6 B. Nouns/Noun Phrases for defining and classifying 19 44 22 38 25 30 12 27 9 23 9 13 24 53 25 45 30 40 16 31 13 28 11 19 IV. Conjunctions A. Additive 4 46 9 43 17 30 9 10 8 19 8 17 B. Alternatives 0 4 0 0 0 0 0 0 0 0 1 1 C. Contrast 0 1 2 0 0 0 0 0 0 2 0 1 D. Others 0 0 0 7 0 0 0 0 0 0 0 0 4 51 11 50 17 30 9 10 8 21 9 19 V. Other Devices related to Defining and Classifying 10 19 4 27 0 13 0 0 1 17 3 2 10 19 4 27 0 13 0 0 1 17 3 2 VI. Descriptive Language (attribution/quality) 0 18 5 14 0 16 0 13 0 8 2 9 0 18 5 14 0 16 0 13 0 8 2 9 Total: 93 349 120 329 142 295 111 145 63 201 84 154 n TO 3 n> CO en o_ fa CO o CO C CO O CO S: o CO " " ^ ° a c n •5' 3 -a o p o. 5' 0X1 es 3 O. T3 O CO Post-R = Post-reading Task Post-R(2) = Post-instruction Task C o O 3 Table 4.12 A summary of the total number of language features found in the post-reading and post-instruction English classification discourse Participant Post-reading Classification Discourse Post-instruction Classification Discourse Result S 73 202 + PA 63 336 + M A 91 200 + A N 131 319 + J 124 198 + SU 60 169 + JO 57 174 + c 48 256 + 54 cr en Participants' English Classification Discourse Post-R Post R(2) Post-R PostR(2) Post-R Post R(2) Post-R Post R(2) Post-R Post R(2) Post-R Post R(2) Language Features S S PA PA MA MA AN AN J J SU SU I Reference (Generic) A. Nouns/ Noun Phrase (Nodes) 35 38 29 109 40 62 42 86 46 52 34 53 B. Noun Phrases (definition) 0 6 2 2 '1 6 0 6 0 1 0 2 C. Examples 8 26 14 31 6 20 41 51 17 24 3 17 43 70 45 142 47 88 83 143 63 77 37 72 II. Transitivity A. Linking Verbs [Relational Process]: (identified - identifier) 6 12 1 17 7 8 2 29 16 16 2 9 B. Linking Verbs [Relational Process] (carrier-attributes) 0 3 0 4 1 1 0 9 0 6 0 6 C. Linking Verbs [Relational Processl: (possessives) 1 8 1 18 4 18 0 15 4 17 1 11 D. Linking Verbs [Existential Processl 1 3 0 2 2 0 3 5 0 1 1 2 E. Action Verbs [Material Processl: Describing behavior 0 12 0 31 1 15 15 28 0 14 0 3 8 38 2 72 15 42 20 86 20 54 4 31 III. Lexis A. Verbs/Verb Phrases for defining and classifying 3 5 4 11 3 10 4 6 4 4 4 5 B. Nouns/Noun Phrases for defining and classifying 11 27 5 32 7 21 5 16 28 28 3 12 14 32 9 43 10 31 9 22 32 32 7 17 TV. Conjunctions A. Additive 7 27 4 36 9 24 19 37 7 12 9 17 B. Alternatives 0 4 0 4 0 0 0 0 0 1 0 2 C. Contrast 0 0 0 1 0 0 0 1 0 0 0 3 D. Others 0 0 0 0 1 1 0 0 0 0 0 0 7 31 4 41 10 25 19 38 7 13 9 22 V. Other Devices related to Defining and Classifying 1 9 3 16 1 5 0 14 2 4 2 12 1 9 3 16 1 5 0 14 2 4 2 12 VI. Descriptive Language (attribution/quality) 0 22 0 22 8 9 0 16 0 18 1 15 0 22 0 22 8 9 0 16 0 18 1 15 Total: 73 202 63 336 91 200 131 319 124 198 60 169 m TO 3 05 c o CO o OS o 3 CL CO a CO C_ *—f CO O 05 3 EL CO ~ O CO O O c CD •a 85 3. -a 05 3 T3 O CO CD 85 a. era 05 3 o. -a o CO 3 CO Post-R = Post-reading Task Post-R(2) = Post-instruction Task C o 5" 3 Table 4.13 (continued) Participants' English Classification Discourse Post-R Post R(2) PostR Post R(2) Language Features JO JO C C I Reference (Generic) A. Nouns/ Noun Phrase (Nodes) 34 48 18 67 B. Noun Phrases (definition) 0 11 1 5 C. Examples 5 22 6 33 39 81 25 105 II. Transitivity A. Linking Verbs [Relational Processl: (identified - identifier) 3 12 3 15 B. Linking Verbs [Relational Processl (carrier-attributes) 0 4 0 9 C. Linking Verbs [Relational Processl: (possessives) 1 12 2 19 D. Linking Verbs [Existential Processl 0 0 0 2 E. Action Verbs [Material Processl: Describing behavior 1 17 7 23 5 45 12 68 III. Lexis A. Verbs/Verb Phrases for defining and classifying 3 1 2 6 B. Nouns/Noun Phrases for defining and classifying 2 9 0 22 5 10 2 28 IV. Conjunctions A. Additive 5 16 5 29 B. Alternatives 0 0 2 0 C. Contrast 0 1 0 0 D. Others 0 0 0 0 5 17 7 29 V. Other Devices related to Defining and Classifying 3 9 1 8 3 9 1 8 VI. Descriptive Language (attribution/quality) 0 12 1 18 0 12 1 18 Total: 57 174 48 256 Post-R = Post-reading Task Post-R(2) = Post-instruction Task 56 4.6.1.2 Quantitative results: Statistical The post-reading Chinese was compared to the post-instruction Chinese using the Wilcoxon Signed Ranks Test on SPSS. The difference was statistically significant. The post-reading English was compared to the post-instruction English using the same test. As with the Chinese, the difference was statistically significant. Table 4.14 The test statistics: Comparison of pre-reading and post-reading Chinese and English classification texts POSTIC - POSTRC POSTIE - POSTRE z -2.201(a) -2.524(a) Asymp. Sig. (1-tailed) .014 .006 a Based on negative ranks b Wilcoxon Signed Ranks Test POSTRC = Post-reading Chinese POSTIC = Post-instructional Chinese POSTRE = Post-reading English POSTIE = Post-instructional English 4.6.1.3 Qualitative results After a one-hour instruction session on the features of classification discourse in the participants' preferred language, their LI (Chinese), the participants showed in the post-instruction task that they were able to produce a more extended classification discourse representing their understanding of the scientific classification of living organisms in Chinese. The following examples in Table 4.15, two texts from one of the participants, SU, will be given to illustrate this. 57 4.15 Examples of SU's texts Text 5 - Participant SU's Post-instruction Chinese Classification Discourse Textual Carrier/Identified Processes existential/relational/material Attributes/Identifier On earth's things ft have thousand million types. However - g f l S * J: they basically can be classified into two big categories. respectively are living things and non-living things ffn while living things can be classified into five big kingdoms respectively are bacteria kingdom, protist kingdom, fungus kingdom, plant kingdom, and animal kingdom Today, I have to introduce is animal kingdom's two big categories -vertebrates and invertebrates invertebrates can be classified into %k&M&m®, mm®®, seven big groups: roundworms, echinoderms, ringed worms flatworms, coelenterates, and arthropods e n They &ft respectively have characteristics &ft respectively have attributes roundworms ft have long body m haven't segments rToa Also most & is Bfft, rounded For example hookworm, ascaris Exactly different's & is ringed worms They similarly ft have long body is. but their body ft have segments 58 Table 4.15 (continued) for example eartfrnwro J For flatworms ft parallel each side for example liver fluke and tapeworms For coelenterates ft is have mourfi's surrounding tentacles mm surface w to sting ce// . _ /(OS in*. radical In addition mollusks's body usually is so/r KAKA-SAI »*f / » rtUJ. Also unsegmented m And echinoderms im like thing its name '*_iffl#r± their body has ffl stings . Qfld radical plans mainly Lastly arthopods ft is animal kingdom's biggest eroup ®{&% occupies have 85%. -ems 77ieir surface m has hard shells Body separated into many parts. ^ ~ — en 77»ey also can be further divided into /owr main groups: myriapods.crustaceans, arachnids, insects For vertebrates also have five main categories These five main categories ft are y?s/>, amphibians, reptiles, birds, and mammals Fish ft w a fype of cold-blooded animal. mm Surface # /ias myx. scales for example *r tove carp, seahorse 59 Table 4.15 (continued) mm Same ft is ft is cold-blooded They and fish biggest differences ft is ftffi is live m # toffcr o« /ana" ffn and & fish jw f/je wafer tEJSI earthworms i t ( f t ) is one of -Attributes/Identifier ft «# They and fish biggest is latter differences is live on land (Post-instruction, Chin, 9.40) Textual);; Carrier/Identified Processes exisbmtial/relational/rnatcnal Attributes/Identifier but reptile is living on ground, (Eng, Post-Instruction, 9.9) and fish is living under water, like snake. (Eng, Post-Instruction, 9.10) but they can live on ground or under the water, like frog and toad. (Eng, Post-Instruction, 9.13) 64 4.7 Possible effects of transfer on students' language production Were there positive and/or negative effects of transfer on students' language production in the tasks? A close look at the classification discourses produced by the participants reveals some significant differences between English and Chinese—what Cummins (1980, 1981) stated as "surface features" distinctive to a language. Given the shared features (noun compounding, nominal groups, nominalization, and grammatical metaphor) used in science registers, however, Halliday also acknowledged the differences between English and Chinese: "Of course English and Chinese use different expressions systems and different grammar" (Halliday & Martin, 1993, p. 124). The comparison of the Chinese and English classification discourses in this study has shown that English and Chinese have distinctive surface features for codifying, extending, and transmitting scientific knowledge. For instance, verb tense and articles are grammatical categories not existing in Chinese while the final particle is not found in English. Also, English allows only one finite verb per clause, but Chinese verbs can be freely chained together. Most distinctively, verbless sentences are sometimes allowed in Chinese. For instance, CARRIER - ATTRIBUTES mmmm-Mffi}. roundworms two sides [are] symmetrical. Body [isJ unsegmented. [# - ft *] [M • ft JR], Body [is] long, appear to be cylindrical, [ I i * - ^ i ] < . And two sides [are] sharp. The verb 'to be', which is used in English to link the carrier to its attributes, is omitted in these Chinese sentences found in the model Chinese classification text. In other words, the attributive relationship shown explicitly by the verb 'to be' in English structures can be less explicitly displayed in Chinese by the order of the words in the sentences, putting the carrier right before its attributes, although the inclusion of the Chinese linking word m (the verb 'to be') is also appropriate. 65 The analysis of the participant's English classification discourse revealed that some students' language production in English had been affected by the structures in Chinese. Table 4.20 offers examples of this. Table 4.20 The influence of Chinese on English a. Sentences with verbs freely chained together S T U D E N T S : Textual Carrier/Identified Processes existential/relational/material Attributes/Identifier and mammals breath have born alive, warm-blooded through lungs mammary glands and cerebral hemispheres. (Eng, Post-Instruction, 2.42) S T U D E N T J : Textual Carrier/Identified Processes existential/relational/material Attributes/Identifier Coelenterates is mostly live has in the ocean, tentacles with sting cells, (Eng, Post-Instruction, 8.5) S T U D E N T A N : Textual Carrier/Identified Processes existential/relational/material Attributes/Identifier Planarians is live in the water; (Eng, Post-Instruction, 5.17) Tapeworms and liver fluke are live on other animals' bodies. (Eng, Post-Instruction, 5.18) 66 b. Verbless sentences STUDEN T SU:. Textual Carrier/Identified Processes existential/relational/material Attributes/Identifier Invertebrate haven't Backbone (Eng, Post-Instruction, 9.20) that different from vertebrate. (Eng, Post-Instruction, 9.21) STUDENT S: Textual Carrier/Identified Processes existential/relational/materia! Attributes/Identifier Flatworms usually live flat, long body, in other animals' body. (Eng, Post-Instruction, 2.21) ringworms segmented, long body (Eng, Post-Instruction. 2.25) myriapods segmented with lots of legs such as centipede and millipede. (Eng, Post-Instruction, 2.34) STUDENT J: Textual Carrier/Identified Processes existential/relational/material Attributes/Identifier Flatworms, flat at the back, two-side reflected, i.e., liver fluke. (Eng, Post-Instruction, 8.7) To realize the knowledge structure of classification and description, the participants need to be socialized into the academic discourse of both English and Chinese and to acquire at least the basic lexical and grammatical features needed to represent the knowledge structures in both English and Chinese. They need to know, for example, how they could link the identified to its identifier, the carrier to its attributes, and so on, in both languages. However, rather than separating content and language by contrasting English and Chinese and discussing how the two languages are distinctive from each other, the analysis of the discourse in this study emphasized how these two resources are being linked to meaning, and used by the participants to realize the same knowledge structures of classification. Although the participants were able to produce a more elaborated classification discourse in both languages, some errors found in their English classification discourse suggested that some participants had negatively transferred the surface features of Chinese to their English classification discourse. 67 4.8 Summary of the findings To sum up the findings and respond to the primary research question directly, it can be said that the participants produced six major written classification discourses at three different stages in this study: pre-reading, post-reading, and post-instruction. The findings suggested that the participants were able to transfer the knowledge structure across their first language, Chinese, and their second language, English. Their written work provided evidence that they had successfully transferred their knowledge about scientific classification of living things learned from the model text of classification in Chinese to their second language, English. However, the comparison between the pre-reading and the post-reading classification, and the comparison between the post-reading and post-instruction task, illustrated that it was not until the participants had received instruction in Chinese about writing classification in Chinese that they were able to produce a more elaborated classification discourse in both languages. Although some participants had showed improvement in writing classifications in both their LI and L2 after reading the model text of classification in Chinese, the results were not significant. The increase/decrease in the total number of features found in the pre-reading and post-reading might be attributed to the difference in the number of categories made between the participants' "common sense" classification and the "scientific" classification of living organisms, although the effect of the model text of classification cannot be entirely denied. In addition, indications were also found in the study that the participants were able to benefit from the instruction in their L I . The analysis of their Chinese and English post-instruction classification discourse suggested that their performance in both their LI (Chinese) and L2 (English) classification discourse had improved in similar ways after receiving instructions about writing classification in Chinese. This supports Halliday's conception of the language of science across languages, which in this case, is across Chinese and English. It may be that since Chinese and English share some common features of science registers, the participants in this study were able to improve in writing classification discourse in both their first and second languages, despite the fact that the instruction about writing classification was given only in one language. Take Participant SU's as an example. His categories of classification in both his Chinese and English pre-reading and post-reading classification discourses were not 68 elaborated with definitions and illustrations. In both of his pre-reading and post-reading classification discourse, he had not given definitions or used linking verbs to link carriers to their attributes. Furthermore, action verbs used to describe the generalized participants' behavior and other descriptive language used to describe factual and precise attribution or quality of the categories were missing in his pre-reading and post-reading classification discourse. In addition, limited examples were given to exemplify the generalized participants. However, after receiving instruction in Chinese about writing classification in Chinese, Participant SU was able to produce a more elaborated classification discourse in both Chinese and English. Examples of most language features commonly used to form classification could be found in his Chinese (LI) and English (L2) post-instruction classification discourse. Table 4.21 summarizes the total number of language features of classification found in Participant SU's LI and L2 classification discourse produced in the three major tasks. Table 4.21 A summary of SU's language features across the three writing tasks Pre-reading (Pre-R) Post-reading (Post-R) Post-instruction (Post-R2) Chinese 67 63 201 English 53 60 169 69 CHAPTER 5: CONCLUSION 5.1 Discussion of findings This chapter will begin by discussing the findings in light of current theories, primarily Cummins's threshold theory and common underlying proficiency, Halliday's conception of science registers across languages, and the potential role of graphics, particularly classification graphics, in bridging language and knowledge. The chapter will then present implications for research and education which stem from this research. It will end with directions for future research. 5.1.1 Cummins's threshold theory and common underlying proficiency The participants in this study seemed to benefit from learning in their LI and demonstrated in the tasks to have the ability to transfer their learning of the scientific classification of living organisms to their L2, as they had all shown in the pre-reading task that they had acquired the basic L2 competency needed to realize the classification knowledge structure in English. The results of this study are broadly consistent with Cummins's claims about the effects of bilingualism. According to his Dual Iceberg Analogy theory, both the LI and L2 of a bilingual operate through the same "Central Operating System" (Baker, 1993) and both languages share a "Common Underlying Proficiency (CUP)" (Cummins, 1981, p. 25). In other words, the thinking skills that the ESL learners acquire through their first language may aid their understanding of the same concept presented in their L2. Yet this study suggests the need to go beyond Cummins' general claim and take account of the specific features of scientific classification discourse across languages. 5.1.2 Halliday's conception of science registers across languages The findings demonstrated that the participants, having learned the formal classification of living organisms in Chinese and having received instruction in the discourse features of classification in Chinese, were able to reproduce their knowledge in both their first language, 70 Chinese, and their second language, English. They were able to form more elaborated classification discourse in both Chinese and English, despite having received instruction about writing classifications only in Chinese. The results of this study shed light on Halliday's conception of the language of science (Halliday & Martin, 1993), suggesting that there may possibly be a "language of science" across Chinese and English. 5.1.3 The potential roles of classification graphics and knowledge structure graphics generally—A means to represent the bilingual's understanding of transfer across languages This section examines the potential productive function of graphics in helping students learn new language and content or transfer knowledge from one language to another. 5.1.3.1 Organize knowledge, and stimulate questioning and reflections on the content knowledge Used as a medium to represent the participants' knowledge of the world, the classification tree has been shown to be a useful tool which can help learners organize and reconstruct their understanding of the content knowledge. Through the expression of their understanding and visualization of the knowledge structure of the text using a classification tree, the participants were engaged in a process of questioning and rethinking—which are both critical thinking skills for promoting students' learning and reconstruction of their knowledge about the world. 5.1.3.2 Represent knowledge of classification and other knowledge structures The comparison of the classification graphics and the classification texts in the pre-reading, the post-reading, and the post-instruction task shows that graphic organizers are a valid means of representing one's knowledge of the world. The classification trees constructed by the participants reflected the same content and levels of classification described in the Chinese and English classification texts. In other words, the use of classification graphics is a potential alternative to the spoken/written classification texts through which students can present their classifications of the world. 71 5.1.3.3 Stimulate negotiation of meaning, and help learners develop and reconstruct their understanding of classification and other knowledge structures The co-construction of the classification tree stimulated learners' rethinking of the content knowledge and promoted negotiation of their interpretation of the texts—a context in which they can construct a common understanding of the texts (knowledge) within their zone of proximal development (Vygotsky, 1978). It encouraged interaction among learners in the co-operative learning task, providing learners with an opportunity for scaffolding their learning and reconstructing their understanding of the content. In other words, the classification graphic was demonstrated to be a mediator which fostered learners' discussion, promoting the scaffolding of knowledge between the learners. 5.1.3.4 Facilitate the transfer of knowledge across the L I and L 2 This study revealed that classification graphics, and possibly other knowledge structure graphics (KS graphics), can help learners both to manipulate the content and to realize their understanding in both Chinese and English. This suggests that knowledge structures do not exist only in the English medium but in both English and Chinese. It is a semantic structure found across languages (Mohan, 1986)—at least across English and Chinese, if not all. The classification tree, which is a graphic representation of the knowledge structure of classification, is thus believed to be an effective tool which bridges the knowledge structures of both languages. It can help to activate learners' knowledge about the semantic relationship and how it is realized in English and Chinese texts. By mediating the transfer of thinking skills and content knowledge across the learners' LI and L2, it is believed that graphic organizers can help learners extend their LI and L2 uses; in other words, they can help maintain and develop the academic use of both the LI and L2, which consequently may help learners achieve a higher threshold where learners can benefit from using both the LI and L2 as resources for learning. This claim is supported by Tang's study (1994a), which demonstrated that "knowledge structures and graphic representations of knowledge structures permit transfer from one language to another" (p. 104). 72 5.1.3.5 A s s i s t the r e a l i z a t i o n o f k n o w l e d g e s t ruc tu res i n L I a n d L2 a c a d e m i c d i s cou r se Classification trees are an effective means of exhibiting the semantic relationships of classification, a major higher-order thinking skill required to achieve academic success across the curriculum. The nodes are formed by nouns/noun phrases, which in the case of this study, are mostly scientific (technical) terms representing different groupings of living organisms. General categories in the classification are always at the top of the graphic, while at the lower level there are sub-categories and specific examples which illustrate the higher-level categories. The lines/arrows in the graphic are meaningful constructions which link and show the relationships between the scientific terms and thus identify each group. Mediated by the same classification tree they constructed in the pre-reading task, the participants were able to transfer their background knowledge (their common sense and/or what they have learned before) from their first language, Chinese, to their second language, English, and use appropriate features of the classification discourse in both languages to represent their knowledge. The student-constructed classification trees, which represented the knowledge structure of classification in Chinese, helped to activate previous understandings about the knowledge structure of classification in Chinese, which included knowledge about the semantic relationship of classification and how the relationship is constructed in English discourse. As Mohan (1989) contended, "knowledge structures are a bridge between the organisation of content knowledge, graphic semiotics, and expository discourse" (p. 105), and "there is considerable evidence that KS graphic representations and other similar graphics help comprehension and subject matter achievement" (Mohan, 1986, p. 125). Participants were able to make corresponding language choices from the two language resources they possessed to represent their understanding about the classification of living organisms in both Chinese and English. Figure 5.1 on the following page displays the roles of classification graphics in aiding the transfer of knowledge across the learner's LI and L2. Figure 5.2 presents the role of KS graphics in assisting the realization of knowledge structures in the learner's LI and L2 and shows how Cummins' language proficiency model and Mohan's Knowledge Framework were re-conceptualized in the study. 73 Figure 5.1 Roles of classification graphics in aiding the transfer of knowledge across languages L1 Knowledge Structures CluMficitiin Principle Hi-St L2 lowlec ructui Sequent* I Choice L1 Text Structured ^Ctassitrcation Graphics,^ (SURFACE LEVEL) LI Discourse *\ spoken written L2 Discourse spoken written U Text Structures 3 Figure 5.2 Re-conceptualizing Cummins' Language Proficiency Model (1980, 1981) and Mohan's Knowledge Framework (1986) 74 5.2 Implications This section will discuss the implications this study has for researchers and educators who are interested in teaching academic language and content in general, and science classification discourse specifically. 5.2.1 Cummins's "Linguistic Interdependent Principle" and goals of ESL programs It is undeniable that ESL programs should aim at helping learners function effectively in their L2 in both social and academic contexts. However, instead of aiming only at developing the learners' L2 (which has a risk of replacing the learners' LI) , the goals of the learning practices in the second language learning programs should not only focus on developing learners' academic uses of the L2 but at the same time concentrate on maintaining and extending the learners' LI language uses. In fact, research (e.g., Peal & Lambert, 1962) has shown that learners benefit more cognitively if both their LI and L2 are adequately developed (Cummins' Threshold Theory). In other words, the learner's LI can have an additive effect on L2 acquisition. In light of the results of this study, it is suggested that parents/schools should promote both LI and L2 language acquisition and language use. Teachers should provide an expanded context to widen their students' acquisition and use of language by learners with rich language learning environments that enable them to experience their LI and L2 in both social and academic situations. Exposure to academic texts in their LI might help ESL learners to develop their Cognitive Academic Language Proficiency (CALP) in their L2. 5.2.2 Halliday and raising students' awareness of the language features of classification in their LI and L 2 The analysis done on the English and Chinese texts in this study have supported Halliday's claim that there are the commonalities in the science registers across languages (Halliday & Martin, 1993; Halliday, 1999). Moreover, the findings of the current research project reinforce the importance of instructing students about the shared features of classification across languages. A greater variety of language devices used to represent 75 knowledge structures should be taught so that learners can better produce academic discourse. However, as there might be potential negative effects of transfer on students' language production in their L2, teachers should also raise students' awareness of the features which distinguish their LI from their L2 with regards to the particular knowledge structure they are teaching. To socialize students into academic discourse, it is essential that more models be provided. Chinese and English classification texts are important models for student learning. However, English texts in some ESL countries are occasionally presented in note form rather than in full sentences (Tong, 2000). Misinterpreting students' failure of comprehension to be merely language specific deficits (Carrell, 1984), many authors of textbooks often break longer texts down into shorter paragraphs (Tong, 2000). Simplifying the texts might actually impede learners' comprehension of the content knowledge, as the semantic relationships is then made covert. This might also limit students' exposure to the target language, which in turn can decelerate their learning of the language features required for representing their academic knowledge. Simplifying the language structures may therefore not be helpful in promoting learning of the L2. In fact, teachers should be aware that there is a need to present knowledge in full sentences and to address the features commonly used across languages to represent various knowledge structures. By displaying the text structure of the knowledge, the semantic relationships of the content knowledge can be made explicit and thus aid the learners' understanding of knowledge structures and content knowledge. Moreover, displaying a range of language features corresponding to the knowledge structures might also help learners develop language which is critical to their academic success, perhaps in both their LI and L2. 5.2.3 Knowledge structure graphics and mediating between students' languages and multimedia To achieve the goals of developing students' higher cognitive academic language in both their LI and L2, assistance should be given to help learners connect and transfer the thinking skills and content knowledge acquired in their first and second languages. KS 76 graphics, which are representations of knowledge structures (Mohan, 1986), are considered to have the potential to mediate the transfer of thinking skills and content knowledge across the learners' LI and L2. First, by activating the learners' background knowledge of the subject and displaying the semantic relationships of the knowledge, KS graphics may assist the comprehension of academic texts in the LI and the L2 (see, for example, Alvermann, 1981; Geva, 1983; Boothby & Alvermann, 1984) and facilitate the transfer of knowledge across the learners' LI and L2. Second, by displaying the text structures (formal schema) of knowledge, KS graphics may also scaffold learners' production of academic texts in their first and second language, helping them make appropriate choices from their language resources to realize the semantic structures of the new knowledge. To acknowledge the important role that KS graphics play in bridging learners' learning of content and language across their LI and L2, teachers should make more extensive use of graphic organizers in class as this will facilitate not only the learner's comprehension of new knowledge but also academic language production and development in both their LI and L2. A fuller range of different functions of graphics should be promoted in actual teaching practice. Also, teachers should help the learners see the value of using graphics as a way to convey ideas beyond the written and spoken mode of language in both their LI and L2, and to regard graphics as a part of the information source (Hurt, 1989). In fact, this process of transfer can be facilitated with the application of multimedia technology. Providing learners with greater flexibility in making revisions, graphic-drawing programs like Paint®, PowerPoint®, and Inspiration® can help learners' rework their knowledge. With the use of such programs, learners can easily make changes to the classifications in the graphic and possibly other KS graphics corresponding to various knowledge structures. Furthermore, as Mohan (1986) stated, as teacher constructed organisers seem to work less well than when students construct their own or complete partially constructed graphics, teachers should also be aware of the need to gradually shift the responsibility to the students and help them make full use of graphics to aid their learning across the curriculum., and ultimately, empower them to apply the uses of graphics organisers as a learning strategy of their own. Only in this way can their learning be transferred beyond immediate lesson. (Mohan, 1986) 77 5.3 Future directions With an understanding of Halliday's view that there are shared features of science register across Chinese and English, I carried out this research to explore the relation between learners' language production of scientific classification discourse in Chinese and English. According to Mohan's (1986) analysis of social practices (see Chapter 2 for details), classification is only one of six core knowledge structures which have been identified; the others are description, principles, sequence, values, and choice, each of which is constructed in discourse by specific features of the lexicogrammar. Future research needs to focus on students' language production of the other knowledge structures across Chinese and English, and possibly across other languages. The results of this study suggest that the skills regarding the semantics of classification which were acquired by second language learners through their first language, Chinese, may help their comprehension of the same concept presented in their second language, English. Moreover, their production of classification discourse in both Chinese and English may have been facilitated through the use of classification graphics. To further explore Halliday's conception of the language of science (Halliday & Martin, 1993), future studies might look at students' language production of classification across other languages and the role which graphics may play in this production. With regards to the various roles that graphics can play in assisting second language learning, this study has integrated the use of classification graphics in the learning tasks. The graphics in these tasks may have been useful in facilitating students' recall and restructuring of information. As an effective means of representing students' knowledge of classification, these graphics also provided additional evidence of the transfer of thinking skills and content knowledge across learners' first and second language. By revealing the structures of classification, the graphics may have also played a facilitative role in aiding students' presentation of knowledge across the two languages in the tasks. However, a complete picture of the role that graphics play in knowledge transfer across languages needs to be subjected to further scrutiny. Moreover, the effects of knowledge structure graphics other than classification graphics should also be explored. 78 REFERENCES Alvermann, D.E. (1981). The compensatory effect of graphic organizers on descriptive text. The Journal of Educational Research, 75, 44-49. Andersson, B.V., & Barnitz, J.G. (1984). Cross-cultural schemata and reading comprehension instruction. Journal of Reading, 28 (2), 102-108. Baker, C. (1993). Foundations of bilingual education and bilingualism. Philadelphia, PA: Multilingual Matters. Boothby, PR. & Alvermann, D.E. (1984). A classroom training study: The effects of graphic organizer instruction on fourth graders' comprehension. Reading World, 23 (4), 325-339. Borg. W., & Gall, M. (1989). Educational research (5th ed.). New York: Longman. Carrell, PL. (1984). Schema theory and ESL reading: Classroom implications and applications. Modern Language Journal, 68 (4), 332-343. Carrell, PL. (1985). Facilitating ESL reading by teaching text structure. TESOL Quarterly, 19 {A), 121-152. Carrell, PL. (1994). Awareness of text structure: Effects on recall. In A.H. Cumming, (Ed.). Bilingual performance in reading and writing (pp. 23-42). Philadelphia, PA: John Benjamins. Christie, F., & Unsworth, L. (2000). Developing socially responsible language research. In L. Unsworth (Ed.), Researching language in schools and communities (pp. 1-26). London, UK: Cassell. Collerson, J. (1994). English grammar: A functional approach. New South Wales: Primary English Teaching Association. Crawford, J. (1998). Ten common fallacies about bilingual education. ERIC Digest. |ED424792| Cummins, J. (1980). The cross-lingual dimensions of language proficiency: Implications for bilingual education and the optimal age issue. TESOL Quarterly, 14 (2), 175-187. Cummins, J. (1981). Empirical and theoretical underpinnings of bilingual education. Journal of Education, 163(1), 16-29. Cummins, J. (1987). Immersion programs: Current issues and future directions. In L.L. Stewin & S-J-H. McCann, (Eds.), Contemporary educational issues: The Canadian mosaic (pp. 192-206). Toronto: Copp Clark Pitman. 79 Derewianka, B. (1991). Exploring how texts work. New South Wales: Primary English Teaching Association. Derewianka, B. (1998). A grammar companion for primary teachers. New South Wales: Primary English Teaching Association. Dunbar, S. (1992). Integrating language and content: A case study. TESL Canada Journal, 70(1), 62-70. Early, M . (1991). Language and content learning K through 12: The Vancouver School Board project. Cross Currents, 18, 179-182, Early, M . , Mohan, B., & Hooper, H. (1989). The Vancouver School Board language and content project. In Esling, J.H. (Ed.), Multicultural education and policy: ESL in the 90s (pp. 107-122). Toronto: O.I.S.E. Geva, E. (1983). Facilitating reading comprehension through flowcharting. Reading Research Quarterly, 18, 384-404. Griffin, C.C., Simmons, D.C., & Kameenui, E.J. (1991). Investigating the effectiveness of graphic organizer instruction of the comprehension and recall of science context by students with learning disabilities. Reading, Writing, and Learning Disabilities, 7, 355-376. Halliday, M . A . K . (1993). The analysis of scientific texts in English and Chinese. In M . A . K . Halliday & J.R. Martin (Eds.), Writing science: Literacy and discursive power (pp. 124-132). London: University of Pittsburgh Press. Halliday, M . A . K . (1994). An introduction to functional grammar (2nd ed.). New York: Arnold. Halliday, M . A . K . (1998). Things and relations: Regrammaticising experience as technical knowledge. In J.R. Martin & R.Veel (Eds.), Reading science: Critical and functional perspectives on discourses of science (pp. 185-235). New York: Routledge. Halliday, M . A . K . (1999). The notion of 'context' in language education. In M . Ghadessy (Ed.), Text and context in functional linguistics (pp. 1-24). Philadelphia: Benjamins. Halliday, M.A.K . , & Martin, J. (1993). Writing science: Literacy and discursive power. Pittsburgh, PA: The University of Pittsburgh Press. Halliday, M.A.I. , & Matthiessen, C.M.I .M. (1999). Construing experience through meaning: A language-based approach to cognition. London: Continuum. Hasan R. (1999). Speaking with reference to context. In Ghadessy M . (Ed.), Text and Context in Functional Linguistics (pp. 219-328) Amsterdam: Benjamins. Huang J. (2003). Activities as a vehicle for linguistic and sociocultural knowledge at the elementary level. Language Teaching Research, 7 (I), 3-33. 80 Hurt, J.A. (1989). A guide for using visuals in teaching culturally diverse children. Tech Trends, 34 (1), 24-26. Hyerle D. (1996). Visual Tools for Constructing Knowledge. Alexandria, VA: Association for Supervision and Curriculum Development. Krashen, S.D. (1996). Under attack: The case against bilingual education. Culver City, C A : Language Education Associates. Kress, G., & van Leeuwen, T. (1996). Reading images: The grammar of visual design. New York, NY: Routledge. Lessow-Hurley, J. (2003). Meeting the needs of second language learners: An educator's guide. Alexandria, VA: Associate for Supervision and Curriculum Development. Liang X . &. Mohan B. (2002). Dilemmas in cooperative learning and academic proficiency in two languages. Manuscript submitted for publication. Longacre, R.E. (1990). Storyline Concerns. Los Angeles: University of California at Los Angeles Department of Linguistics. Longacre, R. E. (1996). The grammar of discourse (2nd ed.). New York and London: Plenum Press. Martin, J. (1992). English Text. Philadelphia, PA: John Benjamins. Martin J. R. (1997). Analysing genre: functional parameters. In F. Christie & J.R. Martin (Eds.), Genre and institutions: Social processes in the workplace and school (pp. 3-39). London: Continuum Mohan, B.A. (1986). Language and content. Reading, M A : Addison-Wesley. Mohan, B. (1989). Knowledge structures and academic discourse. Word, 40 (1-2), 99-114. Mohan, B. (1990). LEP students and the integration of language and content: Knowledge structures and tasks. Proceedings of the first research symposium on limited English proficient students' issues (pp. 113-160). U.S. Department of Education, Office of Bilingual Education Research and Evaluation. |ED 341 260] Mohan, B. (in press). Knowledge structures in social practices. In J. Cummins & C. Davison (Eds.), The Kluwer handbook of English language teaching. London: Kluwer Academic Publishers. Mohan, B., & Huang, J. (2002). Assessing the integration of language and content in a Mandarin as a foreign language classroom. Linguistics and Education, 13 (3), 405-433. Pang, K.C. (1998). t f P ^ ^ S f l l S f e ^ P l [HKCEE Ji-Chu-Sheng-Wu-Xue (Book 1)|. Hong Kong: Oxford University Press. 8 1 Peal, E., & Lambert, W. E.(1962). The relation of bilingualism to intelligence. Psychological Monographs, 76, 1-23. Prince, C. & Lawrence, L. (1994). School readiness and language minority students: Implications of the first national education goal. Washington: NCBE. Short, D.J. (1993). Assessing integrated language and content instruction. TESOL Quarterly, 27(4): 627-656. Shum, L.L. (2004). The effect of environmental factors on bilingualism among Chinese and Korean Americans. New York, NY: Teachers College, Columbia University. Retrieved August 20, 2004, from http://www.tc.columbia.edu/tesolalwebjournal/archives3_l.htm Siren, U. (1991). Minority language transmission in early childhood: Parental intention and language use. Stockholm, The Netherlands: Institute of International Education, Stockholm University Press. Spradley, J.R (1980). Participant observation. New York, NY: Holt, Rinehard, and Winston. Tang, G.M. (1991a). The role and value of graphic representation of knowledge structures in ESL student learning: An ethnographic study. TESL Canada Journal, 9(1), 29-41. Tang, G.M. (1991b). ESL student perception of student-generated graphics as a reading strategy. Reflections on Canadian Literacy, 9(1), 2-9. Tang, G.M. (1992a). Teaching content and English to speakers of other languages in multicultural classrooms. TESOL Journal, 2 (2), 8-12. Tang, G.M. (1992b). The effect of graphic representation of knowledge structures on ESL reading comprehension. Studies in Second Language Acquisition, 14, 177-195. Tang, G.M. (1993). The value of computer graphics in ESL student learning of language and content. Canadian Journal of Educational Communication, 22 (2), 131 -149. Tang, G.M. (1994a). Bridging Chinese-medium and English-medium content-area instruction through cross-cultural graphic representation of knowledge structures. Institute of Language of Educational Journal, 11, 99-118. Tang, G.M. (1994b). Text illustration: A cross-cultural study and its implications for teachers of language minority students. The Journal of Educational Issues of Language Minority Students, 13, 175-194. Tang, G.M. (1997). From graphic literacy across languages to integrating English and content teaching in vocational settings. Hong Kong Journal of Applied Linguistics, 2(1), 97-114. 82 Tong, E. (2000, February). Learning geography through English as a second language: A study on the roles of graphics. Paper presented at the First Annual Graduate Students' Research Conference on Language and Literacy Education Issues. Vancouver, British Columbia. Thompson, E.H. (1993). It's not so hard: Preparing students for reading comprehension. School in the Middle, 2 (3), 33-36. Unsworth, L. (2001). Teaching multiliteracies across the curriculum: Changing contexts of text and image in classroom practice. Buckingham: University Press. Vygotsky, L.S. (1978). Mind in society: The development of higher psychological processes. Cambridge, M A : Harvard University Press. Werner, O., & Schoepfle, G. (1987). Systematic Fieldwork (Vol. 1 & 2). Newbury Park, C A : Sage. Wiley, T.G. (1997). Myths about language diversity and literacy in the United States. ERIC Digest. |ED407881'| 83 APPENDIX A A c o m p a r i s o n o f the l anguage features i n the C h i n e s e a n d E n g l i s h c l a s s i f i c a t i on texts a. Reference (Generic) Model Classification Text in Chinese Model Classification Text in English i) Nouns/ noun phrases (including modifiers) that create nodes (classes) in the classification tree •'everything on earth" "(living) things/ organisms" "non-living things" "the bacteria kingdom" "the protist kingdom" "the fungus kingdom" "the plant kingdom" "the animal kingdom" "vertebrates'" "invertebrates" "coelenterates" "flatworms" "roundworms" "ringed worms'" "mollusks" "echinoderms" "arthropods" "crustaceans" "arachnids" "myriapods" "insects" "fish" "amphibians" "reptiles'* "birds" "mammals" ii) Noun phrases that define generalized participants "animals with backbones" "animals without backbones" "parasites in:other animals" "marine animals with body built on 5-radical plan" "terrestrial animals with long and segmented body" "aquatic cold-blooded vertebrates" "vertebrates with hard, dry, scaly skin" "vertebrates with feathers covering the body" "vertebrates with hairs on their skins" "tt_^±fi<]Kffi (Things on earth/' ft (living things)" "5Eft (non-living things/' "J^££3i f t (bacteria kingdom) " "Bf^^M (protist kingdom) " •'HjgJr?. (fungus kingdom/' "litftJ^ (plant kingdom/' "SDjft-??- (animal kingdom) "-fffttlftft (vertebrates)" • ' « E ^ f £ i | / ) f t (invertebrates)" "SSfllftft (coelenterates)" " E H IS {flatworms) " "£Sia (roundworms) " (ringed worms)" (mollusks) " "filf* r j ^ ^ J mm (all hard shell animals) " "$fc£!8jft (echinoderms)" "gnKftft (arthropods) " " ¥ n & f J ) ^ (crustaceans) " "i&j&Si (arachnids) " "KfeSi (myriapods) " " E U S (insects)" "&(fish)" "MfSBjft (amphibians) " " INSf - f l t * (reptiles)" ",fe (birds) " "n(fi fLftft (mammals) " Wit "IS-ilJr^^ft" (very tiny and we only can use huge microscope then can see's living things) "%k\Hm%tmxmm&±mm m>"(body inside because of metabolism then produce 's wastes) "^^f f f f^I i f t " (have backbone's animals) (along the back's long and hard structures) -&frff tmmw (no backbone's animals) 84 iii) Examples given to exemplify generalized participants "dogs" "cats" "flowers" "trees" "humans'" "bacteria" "molds" "mildews" "spiders" "crabs" "lobsters" "snails" "jellyfish" "worms" "frogs" "salamanders" "sea anemone" "coral" "hydra" "planarians" "tapeworms" "liver fluke" "hookworm" "Ascaris" "earthworm" "sandworm" "leech" "squid" "octopus" "starfish" "sea urchin" "sea cucumber" "wood louse" "water fleas" "scorpion" "mite" "centipedes" "millipedes" "butterfly" "*f£&tSffWt))#J" (live in ocean 's animals) "7K^tfl#3" (aquatic animals) '^M^E^WlW (terrestrial's animals) "i*n jf jLiJ^ '" (cold-blooded animals) -is iK^mm^.m~mmw (transformedfrom water live to land live's kind of animals) •'W^ikWl^Yadapted to temperature cold-blooded animals) in^]W)W(warm-blooded animals) (stones) " "Sfcloudsf (metals)" "#J (dogs)" "IS (cats) " "7E (flowers)" " W * (trees) " " A H (human-beings) " "BW (bacteria)" "SB! (mold" "HM (cars) " "AM (jellyfish)" "$£H (sea anemone) " "MM (coral)" "zkffi. (hydra)" "M£k (planarians) " "r£H (tapeworms) " "flTH (liverfluke)" "f£)la (hookworm) " "tpHJuk (Ascarisf "SEKI (earthworm and sandworm) " "ztc'ijg (leech)" "AJH.& (octopus)" "Skfe. (squid) " (snail)" "tfe (clam) " (squid)" "#51= (starfish) " ")®Slf (sea urchin) " "fS# (sea cucumber) " -m(crab)" (lobster)" "7RSI (waterflea) " 85 "bee" (scorpion) " "beetle" "%B%k (spider) " "dragonfly" "J3i (mite) " "cockroach" "carp" "shark" (centipede)" "^"AEJB (millipede) " "sea horse" "ftllffi (butterfly) " "salamander" "®t> (bee") "lizard" " ¥ H (beetle)" "snake" "4pf!&I (dragonfly) " "turtle" "4IW (cockroach) " "ostrich" "&I& (carp) " "penguin" "robin" (shark) " "pigeon" "ftSJI (seahorse) " "owl" (frog)" "rat" "41^ (toad) " "wolf" "i$i!£ (salamander) " "kangaroo" "W4I (lizard) " "lion" "K (snake) " "M^ (turtle)" "!£,& (ostrich) " ":iiiti (penguin) " "ft]Ell; (robin)" "n§ (pigeon)',' "l&IJJSCow/;" "JH, (ra^" "fg(ca/T "3& (WO/J9 " " S R (kangaroo)" (lion)" b. Transitivity i) Linking verbs [relational processes]: "is/are" "H (verb-to-be) " Linking the subject to its larger group "belong to" "^ (verb-to-be) *' "UfS (include)" "Ivlte (have/occupy about)" "S (belong to)" ii) Linking verbs [relational "are" "(§fc)H (verb-to-be) " processes]: Linking the subject to its "have become" "^Ff^ft (fifon 't/doesn't seem to characteristics (attributes) "don't seem to be" be)" "M (appear to be)" "i& (verb-to-be) " "Rtt fain on/y/* "!15?.(trj) (show/appear)" " ^ (zieerf