Analysis of the Topical Structure of Composition Written by Taiwanese Students

Chia-Ying Chang, National Chung Cheng University, Chiayi County

Abstract: This study attempted to describe and analyze how the rhetoric pattern and topical structure employed by Taiwanese senior-high school students. The research materials were descriptive essays gathered from 100 senior high school students. 96 samples were selected and were given a quantitative analysis first, to compute the overall frequency of occurrence of different rhetoric patterns. Then, 35 samples were selected as high-rated (15 samples) and low-rated essays (20 samples) according to the grades given by two teachers. In the follow-up analysis, 35 written texts were closely examined by utilizing the Topical Structure Analysis proposed by Lautamatti (1987), which examines the semantic relationship between discourse topics and sentence topics and provides insight into the internal coherence in paragraphs. The researcher investigated the types of progressions that are commonly used in the paragraph, and described how these progressions are carried out differently in highrated essays and low-rated essays. The result indicated that students preferred deductive rhetoric pattern to inductive rhetoric pattern. As for the result of topical structure analysis, the findings revealed that low-rated essays often lacked parallel progressions. In addition, more sequential progressions which were indirectly linked to the discourse topic were identified in low-rated essays. Although no salient difference in the frequency of extended parallel progressions employed in high-rated and low-rated essays, the extended parallel progression were most found in conclusion paragraph in high-rated essays. Furthermore, high-rated essays developed fewer topics than low-rated essays so as to maintain the focus of the essay. That is, proficient writers may exhibit better competence in maintaining the topic continuity in their compositions. In sum, the use of thematic progressions is found to have effect on the coherence in writing. The findings of the study can help EFL writing instructors better understand the nature and characteristics of the written text produced by EFL beginning writers. According to the findings, teachers are suggested to investigate distinctive features of rhetoric patterns preferred by students. Moreover, EFL writing instructors are recommend adopting topical structure analysis to analyze students writings. Key words: topical structure analysis, coherence, rhetoric pattern, EFL writing

Introduction
1

Background of the study According to Almaden (2006), a good writing requires logically consistent ideas in order for the writer to compose a coherent piece of discourse that is comprehensible to the intended audience may be (p. 127). However, for decades, instructors and students in Taiwan believe that a well-written composition lies in grammatically accuracy. Likewise, textbooks put major emphasis on teaching sentence-level grammar rather than coherence in a broader sense (Lee, 2002). It is obvious that students focus more on the lexical and sentence levels rather than on the discourse level. In other words, although students are able to produce a grammatically error-free composition, their compositions still lack a coherent structure. Some studies attributed the incoherence to different cultural backgrounds. Namely, under the influence of different cultural and linguistic traditions, learners perceived different rhetoric patterns in writing in their L1 and L2 (Silva, 1992). Contrastive rhetoric researchers have investigated the similarities and differences between Chinese and Westerns rhetoric. Kaplans (1966) indicated the distinctive features between different languages and rhetoric patterns. Anglo-European writers prefer logical argument while Chinese writers had difficulty follow a coherent structure due to the distinct difference between direct linear structure of English writing style and the spiraling pattern of Chinese rhetoric (Kaplan, 1966). In addition to rhetoric pattern, topic continuity and textual coherence are found differently between Western and Chinese rhetoric (Kirkpatrick, 1997; Spencer, Wang & Xing, 2008). To help explain the coherence in writing, Lautamatti (1987) proposed a textlinguistic research method that investigates how sentence topics combine and progress contributing to the coherence of a text. By utilizing Lautamattis (1987) topical structure analysis, this study is expected to identify the incoherence problems existed in compositions written by EFL learners so as to determine the attribution to incoherence. Purpose of the study There are three main purposes of this study. First, by examining the written texts of Taiwanese learners, the current study intends to investigate the preferred rhetoric patterns employed by Taiwanese EFL students. In addition, this research uses a descriptive-qualitative design to investigate how topic flows in learners written text by utilizing Lautamattis (1987) Topical Structure Analysis. Second, the comparison of the high-rated and low-rated essays in terms of thematic progression is expected to provide an insight of teaching certain type of thematic progression to enhance the coherence and fluency of students composition. Third, the finding will propose new pedagogical application in L2 writing, expecting to help teachers and textbook editors deal with the problem of incoherence existed in current Taiwanese EFL writings.
2

Research questions 1. What were the rhetoric patterns employed in composition produced by Taiwanese EFL senior high school students? 2. Is there any difference in terms of thematic progression between high-rated compositions and low-rated compositions? 3. How were the thematic progression proposed by Lautamatti (1987) carried out in the compositions? Significance of the study Although many alternative writing approaches have been applied in real classroom in order to improve the learners writings, textual incoherence remains an unsolved problem encountered by many EFL writers. In spite of the fact that several studies explored ESL/EFL English compositions in light of discourse analysis in Taiwan, most of the subjects that the studies recruit are college-level or graduate students (Chang, 1997; Chi, 2004; You & Joe, 2002). Few studies put attention on senior high school EFL learners. Moreover, text-analysis research drew focus on the argumentative writing or academic writing, different genre such as descriptive writing are rarely discussed in previous research. Judging from previous findings, digression in composition can hamper the logical development of ideas and breaks the overall textual coherence. However, teachers fail to identify the digressions as a main problem in students essay, and few of them can address incoherence problems in a clear manner in responding to students essays. Since Lautimattis (1987) topical structure analysis is suggested as a more satisfying method for analyzing coherence, the current study aims to investigate to identify and explain topical progressions in senior-high school students descriptive compositions in terms of TSA. It is expected that the result and suggestions offered by this study can contribute to writing instruction.

Literature Review
Rhetorical structures in writing According to Kaplan (1966), each culture has its own paragraph orders that conceal an equally unique logic system since writers may be greatly influenced by cultural and language traditions. Since 1960, rhetoric scholars have investigated the similarities and differences between Chinese and Westerns rhetoric. By examining the traditional Chinese rhetoric structures, salient differences are found between Chinese and Western Rhetoric patterns. First distinction is the style of paragraph development. That is, Chinese writers prefer a delayed argument which starts with a broad topic and narrowing it down late in the writing. It is described as an inductive style of paragraph
3

development. On the contrary, Western Writers tend to start with a clear argument and provide supporting ideas in the following part, which is assumed to be a deductive style (Spencer, Wang & Xing, 2008). The second is that English essays adopt great amount of discourse markers to signal the relations between sentences and paragraphs. Yet, Chinese language places emphasis on coherence of meaning rather than coherence of form (Spencer, Wang & Xing, 2008). The third distinctive feature is the frequency of topic transition. That is, Chinese tend to transit from one view of a situation to another. However, Western rhetoric reveals less frequency of transition of views or less topic changing. To conclude the difference in terms of rhetoric features between Chinese and Western rhetoric, three different features appear the most salient, including the paragraph development, conception of coherence, and frequency of topic transition. Paragraph development in English writing According to Wu & Rubin (2000), the placement of thesis statement can be used as a guideline in determining rhetoric features and the development of paragraph as well. Similarly, Kaplan (1966) distinguish English paragraph by identifying the placement of topic statement. One common type of development usually begins with a thesis statement and followed by a series of subdivions of that topic. The other is the reverse procedure, which states a series of examples and then relates the examples into a central statement at the end of the paragraph. The former type represents the deductive reasoning while the latter stands for the inductive reasoning. Likewise, Cahill (2003) defines a deductive essay as initial thesis plus support and an inductive essay as support plus concluding thesis (p.179). Therefore, the placement of thesis statement could be adopted as a means to identify the paragraph development of essay. Coherence in English writing According to Richards et al. (1985), coherence is interpreted as the relations which connect the meanings of utterances in a discourse or of the sentences in a text. Lee (2002) and Johns (1986), likewise, defines coherence as a feature internal to text, which lies beyond the surface structure of text. Lautamatti (1990) refers coherence to the organization of information in discourse and the development of the semantic content. In order to investigate coherence in text, Lautamatti (1987) explores the whole discourse in terms of the relations between discourse topic and subsequent topics. She proposed Topical Structure Analysis (TSA), which is one method writers have adopted to achieve coherence in writing English. TSA has been acknowledged
4

by researchers in determining thematic development in writings. According to Almaden (2006), The use of TSA as device in investigating the progressions in writing has been validated by several researchers as a clear demonstration of style and strategy in linking ideas within paragraphs and reflection of thought progression (P.129).Furthermore, some researchers also indicated that investigation of topic in discourse is a more promising direction for ESL/EFL studies on coherence issue. (Schneider and Connor, 1990). Lautamattis (1987) Topical Structure Analysis (TSA) Topical structure analysis (TSA), developed by Lautamatti (1978, 1987), investigates how sentence topics combine and progress contributing to the coherence of a text. Through topical structure analysis, Lautamatti identifies three possible types of topic progressions, including parallel progression, sequential progression, and extended parallel progression. Simpson (2000) introduced another type of progression known as extended sequential progression. (See example below) The following quoted excerpt from Simpson (2000) shows how thematic progression is achieved. Italics are used to identify initial sentence element (ISE), which refers to the initially placed discourse material in sentences. Underlining is used to signal mood subject or grammatical subject and bold face is for the topical subject.
(1)

For example, one project I set involved the class devising a board game on a nursery rhyme or folk tale for younger children. (2) The class were reasonably enthusiastic about this until they realized that the younger children were fictional, (3) i.e., they would be playing these games with real children apart from each other. (4) I felt a certain amount of shame here, for I realized that the reason there would be no audience was because I had already decided that those games would not be .good enough. For public consumption. (5) I have frequently arranged real audiences for other classes, but only when I have been confident that the finished product would show the class, the school, and most shamefully of all, myself, in a good light. (6) My other error was not to impose a structure to the work of a deadline by which to finish. (7) Because these were low-ability students, my reasoning ran, they would need more time to complete the activity, (8)and in the way of these things, the children simply filled the available time with low- level busy work-colouring in the board, and making the dice and counters, rather than the more challenging activities such as negotiating group responsibilities, discussing the game or devising the rules. (Holden, 1995, cited from Simpson, 2000) In the excerpt provided, Simpson identified the occurrences of progressions based on Lautamattis proposed TSA as follows: (1)Parallel progression: In Clause 2, the noun phrase the class serves as the ISE, the mood subject, and the topical subject. In the following Clause 3, the pronoun they similarly functions as the three sentence elements. They refers to the
5

noun phrase the class in the preceding clause; thus, it is an example of parallel progression. (2)Extended parallel progression: In Clauses 7 and 8, the topical subjects lowability students and the children both pertain back to the topical subject the class in Clause 2. This is an example of extended parallel progression because there are intervening clauses with different topical subjects between Clauses 2 and 7. And in Clauses 7 and 8, the relationship is parallel progression since these are consecutive clauses with the same topical subjects. (3)Sequential progression: This progression is displayed from Clause 3 to Clause 4 and from Clause 4 to Clause 5. The noun games is mentioned in Clause 3, but it is not the topical subject Games is part of the rheme element of Clause 3 and is taken as the theme or topical subject of Clause 4 as noun phrase those games And in Clause 4, the author mentions audience, which is then used in Clause 5 using real audiences as the topical subject. Although Lautamatti suggested only three types of progression, this investigation and that of Simpsons (2002) revealed a fourth type of progression that can be identified as (4) extended sequential progression. Extended sequential progression occurs when the rheme element of a clause is taken up as the theme of a nonconsecutive clause. In the previous example given by Simpson (2000), he claims that the topic of the game is brought up for the first time in the first clause, but not as the topical subject. This is then repeated as the topical subject of clause 4 and a number of clauses intervene between the first mention of the game and the second use as a topical subject, making this an extended sequential progression. (p. 302). According to Simpson (2000), after identifying the topic subjects of each paragraph, the information can be plotted into a table so as to visualize the topical progression of the paragraph. The following table is a sample graph of the sample writing quoted in Holden (1995). As shown in the following graph, the arrow indicates different types of topic progressions. The arrow shift from clause 1 to clause 4 is extended sequential progression while arrow shift from clause 3 to clause 4 indicates sequential progression. Likewise, arrow of clause 4 to clause 5 is also a sequential progression. Therefore, judging from the table, researcher can describe how the topics are shifted and how the paragraph flow and develop. In sum, TSA can be used as a device in determining the flow of the topics and shift of topic so as to examine the coherence in writings.

Figure 1. Simpsons (2000) visualization of topical progression of paragraph Related studies To describe the relationship between coherence and writing quality, Makinen (1992) applied Topical Structure Analysis to short compositions written in English by students learning English as a foreign language. The analyses show that good writers have the ability to develop the topics in their compositions more evenly across several topic levels than mid-level writers and especially the poor writers. Good writers are more homogeneous as a group in handling topics at higher levels than mid-quality and poor writers. The reverse is true of the lowest topical depths. Good writers tend to return to higher topic levels at the end of their compositions more often than the writers in the other categories. Thus, there seems to be a certain degree of relationship between topical structure and writing quality. Similarly, Hoenish (2008) analyzed the topical structure of accomplished essays and compared the results with the topical structure tacitly preferred by assessors in their judgments of student essays. The paper examined whether the same patterns of topical structure that are rewarded by assessors in student writing, as reported by previous studies, are in fact used by professionals writing within a similar genre. The analysis found that the accomplished essays contain similar patterns of topical structure to the high-rated student essays in Schneider and Connor's (1991) study. The accomplished essays, like the high-rated student essays, contain a large proportion of coherence-building sequential progression and a small proportion of parallel progressions. However, some studies yield the opposite findings. The study of Almaden (2006) reveals that students used parallel progressions in their paragraphs and rated as high quality essays. Meanwhile, Witte (1983) found that low-rated essays contain more sequential progressions than high-rated essays. The study of Phuwichit (2004) showed that there is a high frequency of occurrence of extended parallel progression. Parallel progression was found least frequently used. The researchers analysis is that the participants could
7

write well-developed organized essays. The researchers also concluded that the varying result of the study could probably be due to different interpretations and ideas towards writing coherently.

Methodology
The current study intends to investigate how the rhetoric patterns are used by Taiwanese senior high school students. Meanwhile, the research intends to explore the preferred topical structure employed in students descriptive writing, as well as to investigate the differences in terms of thematic progression between high-rated and low-rated essays. In this chapter, the background of the research, including a brief description of subjects, writing task, scoring scheme, raters and the criteria of classifying high and low-rated essays, would be given in the first part. The following sections elaborate the detail of data collections and data analysis. Subjects The study was conducted on four classes from a senior high school in Tainan in December in 2010. The English proficiency of the subjects was general. The subjects were a total 100 third year students, who had learned English for at least five years. They had undergone series of Chinese-English translation training, and had received general instruction of English paragraph structure. The Writing Task Students were asked to write an approximately 120 words descriptive composition in 50-minute class. During the writing session, the students were provided no interruption or guidance other than explanation of the prompt. The title of the writing task was My Best Friend. The title was chosen for it is believed that writings would be easier if beginner writers are given a topic most related to their life experience. Therefore, students were expected to express as much as possible without constraints. Scoring Scheme Since the writing samples would be assessed and be categorized into high-rated essays and low-rated essays, a well established scoring system was adapted to evaluate the coherence in writing. The coherence of each essay was assessed based on the English writing rating scale extracted from College Entrance Examination Center (Lin, S.O., Huang, T.S., Lin, C.I., & Chen, K.T., 1993). Rating scale ranged from poor (0 point)
8

to excellence (4~5 points) in each rubric. The raters were asked to rate solely on the content and organization rubrics since the focus of the current study lay in the topic development and textual organization. Therefore, the highest score of the essay would be 10 while the lowest score is 0. Raters Two experienced senior-high school teachers, who at average have 7- year teaching experience, were recruited in this study for rating the writing samples produced by students. Each essay was scored by the two raters independently, and the final grade was the average of the score given by both raters. If, in very few cases, one grade was given higher than the other by two points, the essay was submitted to the third rater for another scoring. The final grade would be the average of the two similar scores. 3.1.5 Selecting High-Rated and Low-Rated Samples The result of the scores showed that 15 samples scored higher than 8 points while 20 samples were scored under 3 points. Consequently, the high-rated essays, composed of 15 essays, came from those graded 8 and above. On the other hand, the low-rated essays comprised those graded below 3. The Instrument for Analyzing Rhetorical Patterns The writing samples were examined in terms of their placement of thesis statement, and the arrangement of supporting ideas, in order to determine deductive or inductive rhetoric pattern of students essay. The thesis statement is defined as a topic sentence that provides the stance, the belief, or the point of view; or it states the purpose of the essay and conveys the central or main idea of the text (Yang & Cahill, 2008). The current study modified the classification based on the study of Liu (2009) and Petric (2005) concerning to rhetoric pattern. Five rhetoric patterns in total according of the placement of the thesis statement were classified in the following chart. Location of thesis statement Definition 1. Thesis in the first sentence of Topic sentence appears as the first the beginning paragraph (TFB) sentence of the essay. 2. Thesis in the beginning The topic sentence occurs somewhere paragraph (TB) in the introduction paragraph. 3. Thesis in the last sentence of the Topic sentence as the final sentence beginning paragraph (TLB) of the introduction paragraph. 4. Thesis in the middle (TM) Topic sentence was postponed in the body- part of the essay 5. Thesis in the end (TE) A delayed topic sentence appears in the conclusion paragraph. Table 3.1 The coding scheme of rhetoric patterns
9

The Instrument for Analyzing Topical Structure The thematic progression was analyzed according to Topical Structure Analysis proposed by Lautamatti (1987) in order to examine the shift of topics and topic progressions and to investigate the text coherence. One way of conducting TSA is to follow the three steps proposed by Connor and Farmer (1990). The first step is to identify the sentence topics. The second step is to identify the progression type. The last step is to chart the progression of sentence topics. This section described how TSA is conducted by following the three-stage approach devised by Connor and Farmer (1990). Identifying the sentence topic The first step of utilizing TSA is to identify the sentence topic. The topic sentence refers to what the sentence is about. In other words, the main idea of the sentence is topic sentence, which often, but not always, co-occurs with the grammatical subject of the sentence. (Connor & Farmer, 1990). Most of the time, a noun or a noun phrase expressed the topic. The remaining part of the sentence is the comment, which can be explained as what is being said about the topic. The noun that express the sentence topic can occur in many places in a sentence, including beginning, middle, or end of the sentence. Examples excerpted from Lautimattis (1987) studies are presented as below. (1) Newborn infants are completely helpless. (2) Biologists suggest that newborn children are helpless. (3) There are many newborn children who are helpless. (4) Although young animals are certainly at risk, sometimes for weeks or even months after birth, compared with the human infant, they very quickly develop the capacity to fend for themselves. In example (1), the sentence topic is newborn baby and the comment is is completely helpless. The sentence subject newborn infants coincide with the grammatical subject of the sentence. The topical subject newborn children in Example (2) and (3) both appear in the middle of the sentence. The sentence topic of Example (2) is newborn children instead of the grammatical subject Biologists of the main clause. The subject newborn children of the sub-clause served as the sentence topic. Similarly, sentence topic of Example (3) is not the grammatical subject but the discourse topic newborn children, which refers to the idea discussed. In Example (4), the topic of the main clause and that of the sub-clause has the same referent young animals. Therefore, either young animals or they could be sentence topic.
10

Identifying the topical progression After identifying the sentence topic of each sentence, the second step is to identify the topical progression. Lautimatti conducted analysis based on three kinds of progression, including parallel, sequential, and extended parallel. Sample passages showing three types of sentence progression and coherent diagrams excerpted from Connor and Farmer (1990) are presented below. Sample Passages Showing Three Types of Sentence Progressions and Coherence Diagram. Parallel progression SAMPLE1 (1) Chocolates are a national craving. (2) Records show that they are sold in huge quantities 11.2 pounds per capita per year. (3) Designer chocolates often sell for nearly $30/lb. (4) It is obviously that these candies are Americas number one choice. 1. Chocolates 2. they 3. Designer chocolates 4. these candies Sequential progression SAMPLE 2 (1) Computer interviews are used by market researchers to assess product demand. (2) Using these, many different products are analyzed (3) For example, people may be asked about detergents. 1. Computer interviews 2. products 3. detergents SAMPLE 3 (1) I saw a strange man in the street. (2) He was wearing a large badge. (3) The badge was multicolored and gaudy. 1. I 2. He 3. badge SAMPLE 4 (1) School is boring. (2) Many of my friends have motor cycles. (3) As you know, most teenagers have part-time jobs.
11

1. School 2. friends 3. teenagers Extended parallel progression SAMPLE 5 (1) Body language varies from culture to culture. (2) To say yes, Americans nod their heads up and down. (3) Japanese and Italians use the same nod to say no. (4) Body language is an important skill for international managers. 1. Body language 2. Americans 3. Japanese and Italians 4. Body language Table 3.2 Sample passages showing three types of sentence progressions and coherence diagram. (Connor and Farmer, 1990) Based on Lautimatti, parallel progression was where the sentence topics of successive sentences have the same referent. As shown in Sample 1, the topic of each sentence has the same referent chocolate. Sequential progression referred the sentence topic is the rhematic part of the previous sentence. Sample 3 shows a classic case of sequential development, where a comment of a sentence becomes the topic of the nest sentence, whereas Sample 2 is another case of sequential progression. Sample 4 illustrated the sequential progression where the topic of each sentence is unrelated. Sample 5 showed the topical progression of extended parallel. The first three sentences formed sequential progression and the first sentence and the last sentence formed an extended parallel progression. That is, the sentence topic in the first sentence and that in the last sentence had the same referent. Thus, the extended parallel progression is formed. In order to determine the topical progression, a coding guideline for topical structure analysis proposed by Schneider and Conner (1991) was adopted in this study. In the current study, the researcher used the t-units as the unit of analysis and followed Scheidner and Connors (1991) guidelines for coding them. The rationale for using tunits was that they distinguish between simple and compound sentences, allowing clauses separated by coordinating conjunctions to be evaluated as complete topicbearing units. (Hoenish, 2008, p.36). Meanwhile, the criteria of coding three progressions also follow the study of Scheider and Connor (1991). The guidelines
12

were replicated below. Coding Guideline for Topical Structure Analysis T-Units (T) 1. Any independent clause and all its required modifiers. 2. Any non-independent clause punctuated as a sentence (as indicated by end punctuation). 3. Any imperative. Parallel Progression (P) 1. Any sentence topic that exactly repeats, is a pronominal form, or is a synonym of the immediately preceding sentence topic. 2. Any sentence topic that is a singular or plural form of the immediately preceding sentence topic. 3. Any sentence topic that is an affirmative or negative form of the immediately preceding sentence topic (e.g., artists, no artists). 4. Any sentence topic that has the same head noun as the immediately preceding sentence topic (e.g., the ideas of scientists, the ideas of artists; the contributions made by scientists, the contributions made by artists). Sequential Progression (S) 1. Any sentence topic that is different from the immediately preceding sentence topic, that is, not (1)-(4), or P. 2. Any sentence topic in which there is a qualifier that so limits or further specifies an NP that it refers to a different referent (e.g., a nation; a very small, multi-racial nation, referring to two different nations). 3. Any sentence topic that is a derivation of an immediately preceding sentence topic (science, scientists). 4. Any sentence topic that is related to the immediately preceding sentence topic by a part-whole relationship (e.g., these groups, housewives, children, old people). 5. Any sentence topic that repeats a part but not all of an immediately preceding sentence topic (e.g., science and art, science, art). Extended Parallel Progression (Ex) Any sentence topic that is interrupted by at least one sequential topic before it returns to a previous sentence topic. Table 3.3 Schneider and Connor (1991) Coding Guidelines for Topical Structure Analysis Charting the progressions of sentence topics
13

The last step of conducting TSA is to plot those sentence topics onto a graph for a visual presentation of the types of progression employed in the essay. The following text showed how the students essays were coded and how a diagram is plotted. (Phrase in bold refers to sentence topic) Example 1 (1) It has charming eyes, yellow fur and shorty limbs. (2)It is my pet, Bingo, which has accompanied me for ten years. (3)Bingo also is my best friend even though she is a dog. (4) I found her in the park, shaking coldly in the box. (5) When I first saw her at first sight, I then decided to bring it back. (6) Without any doubt, my parents disapprove of it. (7) Eventually, I promised my parents that I would take care of her. (8) Of course, Bingo was my responsibility now. Clause Topical Depth No. 1 2 3 1. it 2. it 3. Bingo 4. I 5. I 6. my parents 7. I 8. Bingo Topic No. 1 1 1 2 2 3 2 1

Example 1contained eight sentences and three sentence topics. Sentence 1 and 2, sentence 2 and 3, sentence 4 and 5 were identified as parallel progression. Sentence 3 and 4, sentence 5 and 6 were identified as sequential progression. Sentence 3 and sentence 8 formed an extended parallel progression. Procedure of Data Collection The following chart shows the procedure of data collection.

14

100 subject were asked to write the essay titled My best friend

4 Essays that were written fewer than 10 clauses were eliminated.

96 essays were collected and were given to two raters.

Essays were first analyzed in terms of the rhetoric pattern.

After the grading process, essays were divided into high-rated and low-rated essays according to the given grades.

15 high-rated essays and 20 low-rated essays were analyzed in terms of topical structure analysis.

15

Table 3.4 Procedure of data collection

Results & Discussion

Rhetoric Pattern of Writing Samples In this section, both quantitative and qualitative analyses are discussed. The quantitative analysis presents the overall frequency of the rhetoric patterns used in students samples. The qualitative analysis describes how the rhetoric patterns were used in students samples. Quantitative Analysis The thesis statement should provide the stance, the belief, or the point of view; or it states the purpose of the essay and conveys the central or main idea of the text. (Yang & Cahill, 2008). Descriptive essay is a wide-spread type that is intended to describe an object, an even, or a person. Descriptive usually follows introduction-bodyconclusion structure. It is important for descriptive writing to include interesting facts in introduction part so that the reader will continue reading the paper. According to many writing guide lines and textbooks, three placement of thesis statement in descriptive essays are suggested. First, a thesis statement is suggested to be included somewhere in the introduction part in order to identify thing or person discussed in the following paragraphs. Second, according to the writing guidelines provided by Roane State Community College, to create a clear pattern of organization in descriptive essays, an introduction should work from general to specific, ending with a thesis sentence. Third, Richard Norquist (1987) in his book Passages: A Writer's Guide indicates that descriptive writing should begin with a topics sentence that identifies the event or person, and explain its significance later on. Similarly, Oshima & Hogue (2007) declare that the topic sentence of a descriptive paragraph should name the topic. That is, a descriptive essay is better start with a clear identification of the topic the writer is going to describe or discuss. In conclusion, it is important for descriptive essays to include a clear topic sentence in the introduction paragraph. As to the placement of the thesis statement, one group of scholars suggested having a thesis statement in the beginning of the introduction paragraph. Another declare the topic sentence should be placed somewhere in the introduction part. The other scholars recommend a general to specific organization and have the topic sentence appear at the end of the introduction paragraph. Three types of thesis placement were found in the current research. Besides, two more different placements of topic sentence were also investigated in students essays. Five rhetoric patterns according of the placement of the thesis statement were classified in the following chart. The first one which thesis statement appears in the beginning sentence is regarded as Thesis in the first sentence of the beginning paragraph (TFB). The second type refers to the
16

thesis statement that is placed somewhere in the beginning paragraph. (TB). The third one which place thesis statement at the end of introduction paragraph is classified as Thesis in the last sentence of the beginning paragraph (TLB). The forth rhetoric pattern refers to a postponed thesis statement in the second paragraph, which place thesis statement in the middle of the essays (TM). The last type refers to those writings with a delayed thesis statement in the end of the essays (TE). In the present essay, rhetoric pattern was indexed by the placement of thesis statement. When the thesis statement appeared in the first paragraph, it was regarded as deductive rhetoric pattern. Similarly to the explanation given by Cahill (2003), a deductive essay means initial thesis plus support and an inductive essay means support plus concluding thesis (Cahill, 2003, p. 179). Therefore, the thesis statement placed in the introduction paragraph, including TFB, TB, and TLB, showed a deductive pattern. Thesis-at-end (TE) pattern reflected an inductive pattern. Thesis-in-middle (TM) pattern combined both deductive and inductive patterns. Table 4.1 illustrated the frequency of rhetoric patterns used in students writings. Table 4.1 The frequency of rhetoric patterns used in students writings Location of thesis statement Total Percentage (%) 1. Thesis in the first sentence 39 41% of the beginning paragraph (TFB) 2. Thesis in the beginning 24 25% paragraph (TB) 3. Thesis in the last sentence 7 7% of the beginning paragraph (TLB) 4. Thesis in the middle (TM) 23 24% 5. Thesis in the end (TE) 3 3% Generally speaking, the findings indicated that the deductive pattern (TFB, TB, TLB) were employed most frequently, followed by non-deductive and non-inductive rhetoric pattern (TM). Inductive pattern (TE) was the least used rhetoric pattern. A more detailed discussion of how the rhetoric patterns were carried out in students writing are presented in the following section. Qualitative Analysis This section discusses five rhetoric patterns, including thesis in the first sentence of the beginning paragraph (TFB), thesis in the beginning paragraph (TB), thesis in the last sentence of beginning paragraph (TL), thesis in the middle (TM), and thesis in the end (TE). The differences of each rhetoric patterns as well as the arrangement of
17

thesis statement are discussed below. 1. Thesis in the first sentence of the beginning paragraph (TFB) The type that thesis statement appears in the first sentence of the beginning paragraph, which is also the introduction of the whole essay, was the most common type used by students. Several kinds of opening sentence were found in these samples. Most students tend to use a direct and explicit way of opening an introduction like the following examples. The following passage is a typical type that straightly identifies the main idea of the essay. My best friend is Ron Artest. Artest and I were known in elementary school. Artest is a good man. His habit is very forthright but he is also impulsive. We all play basketball after school. Artest is very strong and he always wins in basketball game. (Text 1) My best friend is books. Reading books not only increase my knowledge but also my vision. When I was in bad mood, I would read some books and forget those which made me sad and failed. There is a saying goes, He who makes a mistake makes nothing. By reading good books, the more I learned, the more I saw my ignorance. (Text 2) The opening sentence identified the thing or person the author was going to describe in the following passages. Therefore, the opening sentence was meanwhile the thesis statement of the whole essay that conveyed the central idea of the essay. Beside the typical type such as Text 1 and Text 2 that explicitly identified the main topic, other verified ways of opening an introduction were found in other students samples, which are shown as follows. Daphne Wang was my best friend in senior high school. I like her because she had a kind heart to help people and she always had a good performance on her schoolwork. (Text 3) Text 3 used the persons name as the grammatical subject of the opening sentence. Slightly different to the previous examples that start with my best friend, the persons name was introduced at the beginning. The name of the best friend was reinforced in the first place. The main idea was presented directly and explicitly as well. When it comes to my best friend, she is a pretty, charming and active person.
18

Her name is Annie, who accompanies with me and get through many difficult time in my senior high school life. I remember that out first sight was in junior cram school. Her outstanding personality appeals me to meet her. At that time, we were just friend, not best friend. As I entered to senior high school, we met again and became best friend from then on. (Text 4) When it comes to was used often in students samples. It helped the reader to grasp the idea the author tried to convey in that passage. I want to introduce one of my best friends, Hattie. She is really a coolest person I have ever met before. She is good at art and design. Furthermore, she is also a good photographer. She always thinks differently with others. You will amaze what she ever said and done before. (Text 5) Some students tend to use a more colloquial way to start the topic. Essays that write with a colloquial tone tend to consider the reader as a listener. Therefore, the essay would sound more like a letter or a transcript rather than a formal writing. Like Text 5, the personal pronoun you appeared in the last sentence (you will amaze what she ever said and done before) of the paragraph. The stance of the writing has been transferred from the person being described to the reader. The function of placing thesis statement in the beginning sentence is to identify the thing or person later illustrated as well as to help the reader grasp the main idea as soon as possible. It was indicated that most students, 39 out of 96, tend to place thesis statement in the first sentence of the introduction paragraph. 2. Thesis in the beginning paragraph (TB) Thesis in the beginning paragraph refers to the thesis statement was placed somewhere in the beginning paragraph. Notably, it was found that students tend to put thesis statement in the second or third sentence in the beginning paragraph. They were often placed after a given definition of the topic or after the description of the person which would be discussed later. This section presents several texts with thesis statement placed somewhere in the beginning paragraph. Everyone has their friends and one of them is their best friend. Of course, I have my best friend who named Julie. When I met Julie at the first time, I felt she is vivid, cute and optimistic. I like her and view her as my best friend, so does she. Julie does everything a gentle heart and usually gives me some dessert to thank me that I take good care of her. Although we sometimes quarrel with some trifle
19

but we will make peace with each other. Julie is good at Chinese and English, in the contrast, I am cleaver at Chemistry and Math. Therefore, we will teach other which we arent good at. (Text 6) Text 6 is an example of TB. The thesis statement was delivered after giving an opening statement about best friend. A friend can influence one persons life a lot, and it is important for us to meet a good friend in our life time. For me, I had met my best friend in my life, Judy, an energetic girl who always can encourage someone else. With a short hair and black skin, she looks like a borned athlete who is good at any sports. When I got along with her, I always can feel the energy come for her. I will never forget the day we met together. It was the day that we moved to the place which we live now. I was afraid of the new environment because everything was new for me until I met the girl. She quickly relieved my anxiety and took me around to familiar the new place. The girl is Judy, the first friend I met at that day and the most important friend in my life. (Text 7) Text 7 is another type with a short opening about the significance of making good friends. I think that friend is essential in my life. I need friend to help me and give me courage. My best friend is a handsome boy that everyone called him Martin. He is so perfect and popular. He is 180 cm but just 50 kilogram. And he has blue bright eyes, white skin, straight nose and small mouth. No one in school never heard about him. We often play basketball in the gym and do some interesting things our friendship is very staunch like stone that no one can break it. (Text 8) After presenting his point of view about friends, the writer continued to describe the feature of his own best friend. As the proverb goes, A friend in need is a friend indeed. A best friend will give you a hand without hesitation on the right time, giving spiritual support for your injured heart. When it comes to my best friend, needless to say, he must be Jack. Jack, who was one of the most famous students in my class, had not only had great sense of humor but also a thoughtful heart to those who needed help. As far as I concerned, he was just like an idol for me. (Text 9)
20

Text 9 is a common type found in students samples. Proverbs were mostly used as an opening statement. By examining the texts of TB, it was found out that definition and proverbs were commonly used as an opening statement to introduce the following person they were going to discuss. Therefore, a slight difference between TFB and TB is that the opening statement employed in TB. That is, TFB cut into topic straightly and directly while TB indirectly introduces the main idea. Many students employed TB in their writings (25 out of 96). It can be related to the studies that indicate the preference of using proverbs in Chinese students writings. According to Matalene (1985), Chinese students rely on the fixed patterns such as proverbs and idioms. Chinese writers are encouraged to incorporate poems and proverbs in their writings (Wu and Robin, 2000) in order to achieve social harmony. Young (1994) noted that the function of quoting Chinese poetry is to maintain an intimate and interpersonal relation in Chinese society. Moreover, Chen (1994) claims the usage of proverbs is rewarded for increasing the beauty of language. The preference of Chinese students using proverbial reference in Chinese and English writings has been investigated by Wong (1992). He found that the Chinese participants referred to Chinese proverbs frequently in Chinese as well as in English. Therefore, that student using proverbs or general definitions as opening statement is similar to Chinese writings. 3. Thesis in the last sentence of the beginning paragraph (TLB) The last type of deductive rhetoric pattern is to postpone thesis statement at the end of the beginning paragraph. As supported by some scholars, a general to specific introduction is preferred in descriptive essays. That is, the introduction should be ended with a thesis sentence. It is pretty similar to the following texts. He has a vigorous personality. He can bring many laughs to our classroom. He is a pistachio nut. In the class, he can let the boring class be more interesting. In the sad time, he can comfort the badmood friend to make his sadness go away. He is my best friend, John. (Text 10) In Text 10, several sentences were employed to describe the best friend before the author turning to the thesis statement. I will never forget such a valuable experience happening to me while I was a childlike junior high school student. No matter when I recall this unforgettable experience, I still feel embarrassed and cannot help but burst into crying and sorrowful. The memory has remained vivid in my mind because of his handsome
21

face. Tom, whose nickname is Burgerking, is my best friend in my life. (Text 11) Sometimes an experience or a story between the author and the best friend were stated before thesis statement is delivered. Few students employed TLB organization, only 7 out of 96. Samples using TLB is organized in two identical orders. One is to put a description of the person before the main idea. The other is to firstly elaborate the process of how the author and the best friend become friends. The difference among TFB, TB and TLB is the placement in introduction paragraph. All in all, these three patterns are still considered as deductive organization. That is, the thesis statement is stated in the early stage, and followed by detailed supporting ideas in the later paragraphs. 4. Thesis in the middle (TM) TM is the type that cannot be classified as deductive or inductive pattern since the thesis statement is placed in between. It is noteworthy that many students, 24 out of 96, like to postpone their thesis statement to the second paragraph. Examples are shown as follows. Needless to say, people cant live without friends. Friend means the people who can let you open your mind and talk everything to him or her. Since we began to study in kindergarten, we might play with many children. Until we got elder, we may find that only a few people can share your happiness and sorrow, and try to solve any problem together. Those people are called best friends. One of my best friends is called Eason. Eason is not the student who always got good grades in the exam or the most popular guy in our class. Eason is a drawer and an artist, whom many people always tell him be realistic. But in my view, he is the only one person who told me to face my life optimistic. It was a story. I was a student who always has a bad temper, so I usually quarrel with other. One day I fought with my classmates again because he used my pen without my permission. As soon as I started to shout at him, Eason touched my shoulder and told me Look before you leap, try to think if you may still angry when you recall this thing after ten years. Suddenly, I realized how silly I am. From then on, Eason and I became the best friends. His personality can stand my temper when I get angry and complained to him. Sometimes I would protect him when some people laugh his dream. Eason and I could learn more things that I cant found in other people. (Text 12)
22

In the beginning paragraph in Text 12, the definition of best friend was elaborated in the beginning paragraph. Similar to TB we discussed earlier, the definition of best friend was served as an opening statement. The difference lies in the development of the subtopic. That is, TB treats the short opening statement as a subtheme that can be related to the main theme later on whereas TM tends to develop the subtheme into a full length. It was in the cram school that I first saw her. We only had seen each other several times without any interaction. As time went by, we entered the same senior high school unexpectedly. Much to our surprise, when we talked to each other are the first time, we knew that we were in the same class. After entering the same class, we became friends. We accompanied each other to go to the bathroom, chatted about this and that and went shopping together. However, we never told each other everything in our deep hearts. It was not until she encountered some obstacles that we became best friends. As far as I am concerned, she is an essential member in my life. We told each other everything, help each other get over the depression, share our happiness together. She is as generous that she can endure some weakness of mine. One thing is for sure that we have many similar personalities. For example, both of us are pessimistic and sentimental. Because of these reasons, we became best friend undoubtedly and will last our friendship forever. (Text 14) Another type of TM is that the identification of the topic appears at the middle of the second paragraph. Text 14, for example, described how they met each other. However, the main theme of the essay was declared in the final sentence of the second paragraph, followed by the third paragraph illustrating about why they became best friends. The writer used many sentences to develop the procedure of becoming best friends. And the second part of the essay focused on elaborating the significance of the best friend. People benefit from friends a lot. People can share the joy or sadness which their friends, get the assistance when are in need, face to the difficulties with them. Almost everyone has friends and so do I. I have a lot of friends. All of them are kind to me. However, no matter how well the relationship between two guys, there sometimes has some quarrels would happen. But to me, I have a best friend and specialist friend who had never been or showed a bad temper to me. Although he is just passed away
23

when I was ten, he would still live in my mind and would be accompany with me toward the unknown future. He was my dog- Jacob. (Text 15) Text 15 is also another example of TM. It stated the definition of friend in the first paragraph. The second paragraph related the topic with the writer himself, followed by a transition and then turning to the topic sentence. TM is similar to the qi-chenshuan-he essay pattern that is still taught in Taiwanese schools. Chou (1989) proposed two alternative procedures of the development of Chinese rhetoric pattern. One procedure is that introducing the main theme (qi), develop the main theme (cheng), provide anecdote example (zhuang), and conclude (he). The other procedure is to disguise the main theme with a subtheme (qi), develop the subtheme (cheng), introduce the main theme (zhuang) and conclude (he). It is interesting to note that the TM is consistent with the second procedure of essay development. That is, the main theme is introduced until the subtheme is well developed. The examples presented in this section all featured by the development of subtheme. Hence, the main idea is postponed to the second paragraph. Based on the theoretical background, the delay of main theme is probably influenced by the Chinese rhetoric pattern. 5. Thesis in the end (TE) TE is a special rhetoric pattern found in students samples. The thesis statement is postponed in the final paragraph. A delayed thesis statement refers to a more inductive style of rhetoric. That is, many supporting ideas are given earlier than approaching a thesis or theme. Only three out of 96 in the current study used this pattern. The following are the detailed discussion of the rhetoric pattern. She is graceful girl and know lots of things what she learn from extra study. She doesnt like to talk in public, but she can easily express her thought in logic and always be trusted by others. The most special temperament she interests me is the view point she has to her life. Every time when I was in the bad mood she not only gives me a warm hug or listen what make me being upset but cheer me up by sharing some optimistic idea that always enlarge my eyesight seeing the life, the world we live. To my surprise, in my memory, we dont have any quarrels with each other. We always have the opposite thoughts but we will try to understand what the other think, we share our idea and find creation in our conversation. Both of us learn a lot from each other. In the special day like birthday or Christmas, we write long letters to each other to appreciate that we can have such a great friendship. She is my best friend because I know she knows me, will give me a hand
24

when I need someone to accompany. We know we have each other and most important is we know to give others enough space. She is really my best friend. (Text 16) Text 16 is noteworthy that it started without any topic sentence that could clearly identify what is been talking about in the essay. The first and second paragraph failed to directly relate the title with its text. A clear and direct relevance to the title my best friend was given until the final paragraph. A delayed thesis statement is hard for the reader to figure out the focus of the writing without knowing the title. That is, it was in the last paragraph that the authors best friend was identified. I used to be alone when I was in junior high school. I was shy and quiet then so I didnt have any friend. Until one day, she came to me. Her name is Grace. She is a very outgoing and energetic girl. Our first conversation is unforgettable. She said hi, my name is Grace. Since this moment you are my friend. I burst out a laughter as soon as she finished what was going to say. She laughed at me, too. Since then, no matter where we went or what we did, we always be together. We shared out little secrets to each other. If we had something unhappy, we would cheered us up. Time passed so quickly, three years in junior high were about to finish. Both of us were very sad to say goodbye. We will always be in touch. Our friendship would last forever. Grace will be my best friend forever. (Text 17) Text 17 has similarity with the previous one. The paragraph started with description of the person, without relating to the topic of the writing. The distinctive feature of TE is that a clear relevance to the topic is placed in the final paragraph. Consequently, the reader would be confused if the title has not been given in the first place. My best friend cant talk, but it can make me warm every day except rainy day or cloudy. Sometimes I feel bad and want to cry, it always gives me some power and encouragement. Then I feel better and smile again. My best friend is shining every day. Its body has seven kinds of colors. In fact, I cant see it at night because its another friend will come out. That is to say, that is like that we both go to bed at night. When I get up and go to school, it follows me and smiles to me. Many peoples friends can talk, but my best friend cant talk. In my opinion, its Ok because we talk every time and everywhere by using our mind. I think using our mind to talk to each other is better than using mouth. And my best friend
25

also agree it. Now do you know who is my best friend? Thats right, my best friend is the sun. The only sun gives me the only wish. And my best friend gives some powers to realize my dream. (Text 18) The last TE sample is rather interesting. The writer first elaborated many features of her best friend. In the end, she asked the reader to guess the best friend she referred to. Namely, the whole essay was like a puzzle which is comprised by many hints about the best friend, and the answer was given at the end of the essay. TE is a special pattern found in students essay. TE, meanwhile, is an inductive rhetoric pattern which postpones thesis statement till the last paragraph. The current study indicated that the students tend to use a deductive rhetoric pattern in descriptive writings. That is, the main theme of the essay is often illustrated in the early stage of the essay. Topical Structure of Writing Samples In this section, the high-rated and low-rated essays are discussed in terms of thematic progressions. The first part presentes the quantitative analysis of topical progressions employed in high- and low-rated essays respectively. The second part discusses the differences of using topical progressions between high-rated and low-rated essays. Quantitative Analysis The frequency of high-rated essays was shown in Table 4.2.1. The total T-unit of highrated essays was 236. The topic depth refers to the number of different topics introduced in an essay. In high-rated essays, 83 topics were introduced. In other words, about 35 % topics were introduced in high-rated essays. Therefore, the larger the figure of topic depth was, the more topics were introduced. The parallel progression refers to the frequency that parallel progressions appear in high-rated essays. Hence, there were 95 parallel progressions occur in total 236 t-units of highrated essays. That is, the occurrence of parallel progression was around 40 % in highrated essays. The occurrence of sequential progressions in high-rated essays was about 34% whereas the occurrence of extended parallel progression was about 20%. Table 4.2.1 Frequency of Topical Progression in High-rated Essays Total Percentage (Each progression type/T-unit) (%) T-unit 236 --Topic Depth 83 35.16
26

Parallel Progression 95 40.25 Sequential Progression 80 33.89 Extended Parallel Progression 48 20.33 The frequency of topical progression in low-rated essays was illustrated below. The tunits in low-rated essays were 251 in total. The topics introduced in low-rated essays were around 44%. The occurrence of parallel progression was 35%. The occurrences of sequential and extended parallel progression were 38% and 20% respectively. Table 4.2.2 Frequency of Topical Progression in Low-rated Essays Total Percentage (Each progression type/T-unit) (%) T-unit 251 --Topic Depth 110 43.82 Parallel Progression 88 35.05 Sequential Progression 95 37.84 Extended Parallel Progression 51 20.31 The following table showed the comparison of high-rated and low-rated essays. In terms of topic depth, low-rated essays contained more topics introduced in essays. According to the statistics, high-rated essays used parallel progressions more frequently while low-rated essays contained more sequential progressions. Furthermore, no significant difference lay in the occurrence of extended parallel progression between high- and low-rated essays. Table 4.2.3 Comparison of Topical Progression between High-rated Essays and Lowrated Essays High-rated essays Low-rated essays Topic Depth 35.16 43.82 Parallel 40.25 35.05 Progression Sequential 33.89 37.84 Progression Extended Parallel 20.33 20.31 Progression Qualitative Analysis This section presents the detailed discussion of parallel, sequential and extended parallel progressions employed in students essays. The writing samples that best exemplifies topical progressions are presented in this section. Meanwhile, some diagrams of topical progressions that plotted topical subjects into a table are presented when necessary. The function of the diagram is to help explain the development of thematic progressions.
27

Detailed discussion of parallel progression According to Schneider & Connor (1991), the function of parallel progression is to reinforce the idea on the readers mind. In the parallel progressions, repetition of key words and phrases are used to develop the ideas. In the current study, the use of the parallel progression was more frequent in high-rated essays than that in low-rated essays. Moreover, the researcher found several differences in parallel topics between high-rated essays and low-rated ones. The differences can be classified in terms of the use of pronouns, focus of the essay, and the use of subjects. Use of pronouns The following sample is extracted from high-rated samples.
(1)

Friend is a person who gives you a hand when you are in need, a person that connects strong bonds with you. (2) Friends are the ones you can laugh with and share the ups and down with. (3) Daisy, is a friend of mine in junior high school. (4) Shes a considerate and magnanimous person. (5) She is also an outstanding student in class. (6) Every time when I has problems in my school work, she helps me solve them. (7) Whenever problems depressed me, shell always be beside me at no times. (8) As quarrels occurs, she s always the first person to analyze the factors that lead to this fight. (9) As time passed by, these quarrels strengthened the bond between us. (10) Our friendship continues even though we are now indifferent high-schools. (11) For example, we went to China together, go hiking on the weekends, and shopping when we have free time. (12) She is truly my best friend forever. (13) A friend in need is a friend indeed. (14) A friend is a person who doesnt abandon you when youre met with difficulties. (15) A friend who doesnt flatter around you would say the reality beside you. (16) Upon all, Daisy is a true friend that I deeply love, and nothing can break us apart. (Sample H2) 1 2 3 4 5 1. friend 1 2. friends 1 3. Daisy 2 4. she 2 5. she 2 6. she 2 7. she 2 8. she 2 9. these 3 quarrels
28

10. 11. 12. 13. 14. 15. 16. she

our friendship we a friend a friend a friend Daisy

4 2 5 1 1 1 2

The first parallel progression was identified between sentence 1 and sentence 2, the second parallel progression to sixth parallel progression were from sentence 3 to sentence 8 while sentence 13 to sentence 15 formed another two parallel progressions. This sample contained many parallel progressions. The name of best friend was mentioned again at the end of the essay. In other words, though the pronoun she was used consistently, the author mentioned the name of the best friend in conclusion part to recall the readers mind. Similar to this recalling skill, the following sample is another example of parallel progression, which mentioned the name of best friend back and forth to reinforce the central idea of the essay.
(1)

It has charming eyes, yellow fur and shorty limbs. (2)It is my pet, Bingo, which has accompanied me for ten years. (3)Bingo also is my best friend even though she is a dog. (4) I found her in the park, shaking coldly in the box. (5) When I first saw her at first sight, I then decided to bring it back. (6) Without any doubt, my parents disapprove of it. (7) Eventually, I promised my parents that I would take care of her. (8) Of course, Bingo was my responsibility now. (9) She is smart and what I teach her she would remember it and wouldnt make the same mistakes. (10) When I was desperate, she will stay alone with me with no reasons. (11) Whats more, even though she does not understand our language, when I talk to her, she will be very concentrated. (12) Ten years has passed, the friendship between us doesnt disappear. (13) Instead, it has become more stronger. (14) We pass through many difficulties together and share happiness. (15) Bingo is not only my pet now but my best friend. (Sample H1) 1 2 3 4 5 9. it 1 10. it 1 11. Bingo 1 12. I 2 13. I 2 14. my 3 parents
29

15. 16. 17. 18. 19. 20. 21. 22. 23.

I Bingo she she she the friendship it we Bingo

2 1 1 1 1 4 4 5 1

It is notable that the name of the best friend, Bingo, was kept mentioned once in a while throughout the essay. By means of repeating key phrases, the name of the best friend as well as the main topic was reinforced in the essay. A strong impression may be left in the readers mind. However, when compared to high-rated essays, low-rated essays tend to use pronouns solely throughout the essay even if the name of best friend has been mentioned previously. Like the following two examples extracted from low-rated essays, the author tend to use pronouns to refer to the best friend. Nevertheless, the name of the best friend, which has been mentioned previously, failed to appear at the end of the essay. Thus, the impression in the readers mind may not be as strong as utilizing repetition of the key words. As shown in Sample L1, it was used to refer to the best friend even though the name of the best friend is already mentioned in sentence3. At the conclusion part, instead of referring back to the best friend, the writer further brought another idea. Hence, it was easy for readers to lose focus of the essay. Similarly, Sample L2 used pronoun he all the way to refer to the best friend; yet, the name of the best friend, which has been mentioned in sentence 6, never occurred again. Instead, the topic of the essay was ended with we (sentence10 and sentence11). The elaboration of the topic was progressed into a different direction. In other words, the focus of the essay had been shifted from the best friend to we, which caused the impression that the essay has not been finished and could be further developed on the topic we.
(1)

Everyone has a best friend in his or her life. (2) Needless to say, I also have my best friend. (3) It is a puppy named Kiky. (4) It was a newborn stray dog, and was picked up by my parents. (5) It looked like a black fur ball with brown fur on its head and feet. (6) After getting along with it, we had closed to each other. (7) After I returned home from the school, we always played together in the park or share my emotions with it. (8) However, with the passing of the time, it got older and older, and didnt have energy as before. (9) My parents prepared
30

a comfort place for it to live out its life. (10) One day, it ran out from our house and never return. (11) My parents told me that it would die outside because of a legend that dogs would far away from their owner before they died. (12) In spite of sadness, I would treasure the time that we stayed together. (Sample L1) I have many friends, but know my heart is very little. (6) My best friend is Tony, who is a nice guy. (7) When I want to do something bad, he will advise me to think before I act. (8) And when my mood is bad, he will care about me. (9) Because my grade is bad, so he often asks me to study with him. (10) Recall to that all things, he is really my best friend because he cares about me and often do something good for me. (11) And we communicate now. (12) When we each other have hard time, we will help each other. (Sample L2) Both sample L1 and Sample L2 failed to recall the name of the best friend at the end of the essay. Consequently, readers attention was also distracted from the focus of the essay. Examining from high-rated essays, the researcher found that the recall of the name of best friend is likely to leave the reader a stronger impression of what the essays is talking about. In sum, the difference of using pronouns may affect readers impression of the essay. In other words, appropriate repetition of key words can leave reader a clear picture of the main idea. Moreover, the recall of key words at the end of the essay can trace back to the main idea to remind readers the focus of the essay. The relevance of parallel topics The consistency of parallel topics can not only leave a clear impression of the essay but to develop the main idea. By means of repeating the key words, the central idea of the essay can be maintained and stay focused. Meanwhile, readers can easily identify the main idea the writer is trying to convey. It is observed that high-rated essays had competence in using parallel topics to maintain the focus of the essay. The following example is also extracted from high-rated samples.
(1) (5)

When I graduated from junior high school, I entered an ideal senior high school, beginning my thrilling and interesting life in school. (2) Two years ago, I knew a girl. (3) At that time, she was look austere that I felt she was hard to deal with. (4) To my surprise, one day I forgot to bring text book, she found my hardship and gave me help in no time. (5) From then on, we became best friend and always study together until now.
31

(6)

We have many same personality and perspectives. (7) For instance, we are all diligent in study and we like to listen to music and play the piano. (8) We also like frantically to see the comedy movies. (9) Because of these reasons, we became best friend. (10) Our precious memories strengthen our friendship. (11) For example, we once took a trip to Sun Moon Lake together. (12) We took many valuable pictures and had a lot of fun. (13) The entrance exam of university is in the corner, we make our mind to enter an ideal university. (14) When we accomplish our goal, we will learn a new language, French. (15) This imperishable friendship will last forever because its always a chance of lifetime to make a such good friend. (Sample H13) 1 2 3 4 5 1. I 1 2. I 1 3. she 2 4. she 2 5. we 3 6. we 3 7. we 3 8. we 3 9. we 3 10. Our 4 precious memories 11. we 3 12. we 3 13. we 3 14. we 3 15. friendship 5 In this essay, the author first introduces when and how she met her best friend. From sentence 5, we was served as the topical subject which was also carried through sentence 6 until sentence 9. In sentence 10, the topic was developed with our precious memory, which was indirectly related to we, was carried out as a sequential progression. After that, the topic was brought back to we until sentence 14. The writer used we to build an impressive relationship between the writer himself and his best friend. Hence, the idea of the relationship between the writer and his best friend was reinforced throughout the essay. By examining the TSA plotted table, it is easy to tell that the focus of the essay was about we, that is, the author and his/her best friend. This showed that parallel progression was used mainly in this essay. Employing parallel progression in the paragraph is advantageous because this results in strong topic focus. High-rated
32

essays exhibited a great deal of parallel progressions to string ideas closely together. In addition to maintain the focus o the essay, high-rated essays also exhibited great relevance between each parallel topics. As shown in the following paragraphs.
(1)

Friend is the best gist which God give us. (2) I think it describes very well. (3) Jessica, who is my best friend, is a optimistic, generous and thoughtful girl. (4) Because her company, I have many happy times. (5) Jessica is very optimistic about everything. (6) She always brings a big smile on her face. (7) When she faces difficulties, she always looks things on the bright side. (8) I remember one time I felt sad on my bad grade on test, she told me to cheer up and get good grade next time. (9) Her encouragement became confident of myself. (10) Jessica is very generous to people. (11) When she has cookies or candies, she always gives them to others. (12) And when someone needs help, she also help them without thinking. (13) Jessica is also a thoughtful girl. (14) When I sad, she always on my side and company me. (15) Sometimes, she would even give me a big hug and make me feel very sweet. (16) To sum up, Jessica is the best friend in my life. (17) I think it is so lucky to have such a good person to be my friend. (Sample H14) 1 2 3 4 5 6 1. friend 1 2. I 2 3. Jessica 3 4. her 4 company 5. Jessica 3 6. she 3 7. she 3 8. she 3 9. her 5 encouragement 10. Jessica 3 11. she 3 12. she 3 13. Jessica 3 14. she 3 15. she 3 16. Jessica 3 17. to have 6 From second paragraph of the essay, the author introduced Jessica as the topical subject of clause 3. In clause 5, Jessica served as the topical subject which was
33

carried through clause 6 until clause 8 using the pronoun she. In the following paragraphs, Jessica served as the topical subject which was parallel to the topical subject in clause 11 to clasue16. To examine the essay closely, the topic sentence of the essay was in clause 3, which described the best friends as an optimistic, generous and thoughtful girl. The following paragraphs are the supporting ideas, such as clauses 5, 10 and 13, further developed the main idea. Thematic progression was done through repeating the parallel topics. Meanwhile, the supporting ideas were closely related to the main idea and the essay was developed based on the main topic. However, low-rated essays displayed the problem of digression from the main idea. As shown in the following sample extracted from low-rated essay, every sentence cannot be connected and related even though the parallel topics were carried out through the essay.
(1)

My best friend is very strange. (2) He is round, and a little fat. (3) He is a basketball. (4) I met him when he came to my twelve old birthday party. (5) From that day, I played with him every day. (6) When I was delight, I could share my happiness with it. (7) When I was sad, he could cheer me. (8) But he has become older and older now. (9) He was too old to run fast and jump high. (10) Despite he cant play with me, I still like him. (11) He is my best friend forever no matter how old he is. (Sample L10) 1 2 1. my best friend 1 2. he 1 3. he 1 4. he 1 5. I 2 6. I 2 7. he 1 8. he 1 9. he 1 10. he 1 11. my best friend 1 My best friend served as the topical subject of the first sentence was still in reference to the topical subject of sentence2. From sentence 3 to sentence 4 the author used the pronoun he to refered to my best friend. Despite the consistent parallel topics, each sentence was not closely related to the theme of the essay. That is, from sentence 1 to sentence 3, the writer developed the idea that my best friend is strange into he is round and a little fat. Nevertheless, these sentences were not closely related since the writer did not provide the connection between the two such as
34

explaining the reason why strange is related to its appearance. Moreover, the writer explicitly identified the best friend as a basketball in sentence 3. Still. The sentence had no relation to the previous sentences. That is, the flow of the ideas was awkward. At the final part of the essay, the writer mentioned the idea that the ball is too old to play with. The theme of the essay had been digressed to another direction. The whole essay lacked detailed information to connect each sentence. Thus, the whole essay seems reckless and failed to maintain the theme of the essay. In short, though parallel progressions are carried out both in high-rated and low-rated essays, the relevance of parallel topics are apparently different. Detailed discussion of sequential progression As mentioned in the previous section, low-rated essays contained more sequential progression than high-rated essay do. However, the use of sequential progression was found different between high-rated and low-rated essays. According to Schieider & Connor (1991), sequential progression can be further classified as directly related, indirectly related and unrelated topics. Directly related sequential progression include the topic-comment relation, which means the comment mentioned in previous sentence become the topic of the following sentence. Moreover, word deviation (science, scientists) and part-whole relation (these groups, children and old people) are considered as related sequential topics. Indirectly related topics are related by semantic set. For example, relation between scientists and their inventions. Unrelated sequential progressions refer to those topics that neither relates to previous topic nor to the main topic of the essay. It was found out that high-rated essays contained more directly related sequential topics. Yet, low-rated essays contained more unrelated or indirectly related sequential topics. It has been calculated by the researcher that highrated contained 30 directly related sequential topics out of 80 sequential progressions while low-rated contained merely 18 directly related sequential topics out of 95 sequential progressions. Specific examples and comparison are displayed as follows. Directly related sequential progression (topic-comment relation) The following texts are paragraphs extracted from high-rated samples. Since highrated essays employed few sequential progressions, the complete samples would not be presented but short passages that contained sequential progressions. The following text is the first two paragraphs of one high-rated sample. Sequential progression occurred between clause 2 and clause 3. It is a topic-comment relation since a kind heart in clause2 stands for rheme and then became the theme in the next clause. Moreover, at the end of the paragraph, a nice heart occurred again to trace back to the main idea of the paragraph.
35

(1)

Daphne Wang was my best friend in senior high school. (2) She had a kind heart to help people and she always had a good performance on her schoolwork. (3) A friendly heart made people like you. (4) Daphne was careful to every events and every person. (5) Once a little girl lost her way to home, Daphne gave her a big hand to find her way home. (6) According to this incident, I think Daphne is with a nice heart. (Sample H7) Topic-comment relation occurred frequently in high-rated essays. The following texts are the paragraphs that contain directly related sequential topics in terms of topiccomment relation. One day, Timmy disappeared as I didnt close the door. (8) I blamed myself for being so careless. (9) Timmy would not be lost if I take good care of him. (10) After researching for several hours, I still couldnt find him. (11) I was very sad and afraid of not to seeing it again. (12) Fortunately, Timmy appeared at the best moment and I was so glad to hug it for a long time. (13) Though we are not the same species, Timmy and I will have built up a friendship. (14) The friendship is the thing I treasure forever. (Sample H5) 7. Timmy 1 8. I 4 9. Timmy 1 10. I 4 11. I 4 12. Timmy 1 13. we 5 14. friendship 6 The preceding text is passages extracted from one high-rated sample. Three sequential progressions were identified between sentence 7 and 8, sentence 9 and 10, sentence 12 and 13. The last sequential progression appeared between sentence 13 and sentence 14. Friendship was introduced as rheme in sentence 13 and became theme in sentence 14. It is noteworthy that from sentence 7 to sentence 12, the topical subjects I and Timmy were used alternatively, which developed a story happened between the author and Timmy. Moreover, at sentence 13, the topic subject Timmy and I successfully connected the separate pieces into a whole story. Though many sequential progressions were used here, the main topic of this essay had not been digressed. In other words, the topic beat around the unforgettable experience happened between the author and the best friend. There is another example of topic-comment relation.
36
(7)

(3)

Daisy, is a friend of mine in junior high school. (4) Shes a considerate and magnanimous person. (5) She is also an outstanding student in class. (6) Every time when I has problems in my school work, she helps me solve them. (7) Whenever problems depressed me, shell always be beside me at no times. (8) As quarrels occurs, she s always the first person to analyze the factors that lead to this fight. (9) As time passed by, these quarrels strengthened the bond between us. (10) Our friendship continues even though we are now indifferent high-schools. (11) For example, we went to China together, go hiking on the weekends, and shopping when we have free time. (12) She is truly my best friend forever. (Sample H2) 3. Daisy 2 4. she 2 5. she 2 6. she 2 7. she 2 8. she 2 9. these 3 quarrels 10. our 4 friendship 11. we 5 12. she 2 It is apparently that three sequential progressions occured between sentence8 and9, sentence 9 and 10, sentence10and 11. quarrels which was mentioned in the preceding sentence became the theme in sentence 9. Meanwhile, the bond which was rheme in sentence 9 became the theme our friendship in senetnce10. It is apparently that each sentence had direct relation so as to construct a coherent and intact essay. Direct related Sequential progression (word-division) According to Schineider and Connor (1991), word division is considered as one of the directly related sequential progression. For example, the sequential relation between science and scientist is directly related. The following text is a word-division type of sequential progression.
(1)

People say that A best friend is a color that brightens up your life. (2) Undoubtedly, when you get into trouble or sometimes in a bad mood, your best friend will always stay with you. (3) My best friend, Young, showed up when I was 13. (4) She was my new classmate in
37

junior high school first grade, sitting next to me. (5) Just the opposite to me, Young is a gentle and quiet girl, always with smile on her face. (6) Suddenly, we became the greatest friend. (Sample H3) 1. A best 1 friend 2. Your 2 best friend 3. My best 3 friend 4. She 3 5. Young 3 6. We 4 Three sequential progressions were identified between sentence1 and 2, sentence 2 and 3, sentence 5 and 6. Word-division occurred as a best friend, your best friend and my best friend. It seems like the topic was arranged from general to specific. Indirectly related sequential progression Sequential progression contained not only directly related topics but also indirectly related and unrelated topics. Low-rated essays though employed more sequential progressions than high-rated essays do, indirectly related and unrelated topics present high occurrence in low-rated essays. The following two samples were low-rated essays that contain the most sequential progressions. I cant live happily without my best friend, Puni. (2) In fact, hes just a puppy, but full of great energy. (3) We met each other in a rainy day. (4) He was afraid prior to my houses door when I go back to home from school. (5) His pity eyes touch my heart so that I decided to ask my grandpa for feed him. (6) But my grandpa didnt want to do that. (7) During 30 minutes, I never touch my grandpa whose heart is like a stone. (8) The puppy was quake with the cold wind. (9) In that time, my grandma came back become its hero. (Sample L12) 1 2 3 4 5 6 1. my best 1 friend 2. he 1 3. we 2 4. he 1 5. his eyes 3 6. my 4 grandpa 7. I 5 8. the 1
38
(1)

puppy 9. my grandma 6

The organization was similar to the previous example (Sample H5) in high-rated essay. Both authors expressed an experience and a story in the second paragraph. However, the author of low-rated essay used many sentences to describe the persuasion happened between the author and his grandpa, instead of describing the story between the best friend and the writer himself. 5 Sequential progressions occurred between sentence 2 and 3, sentence 4 and 5, sentence 5 and 6, sentence 6 and 7, sentence 8 and 9. The topics shifted quickly and were not directly related. For example, the topical subject he in sentence 4 and the topical subject his pity eyes in sentence 5 was a typical example of indirectly related sequential topics. In addition, sentence 7 (I never touch my grandpa whose heart is like a stone) had little reference to the next sentence (the puppy was quake with the cold wind). When readers attention had been digressed to the puppy, new information in sentence 9 (in that time, my grandma came back become its hero) was introduced. From this evidence, new information was introduced while the given information has not been well explained. The given-new information that has not been arranged and organized would lead to unrelated sentence pieces. Unrelated sequential topics failed to connect the topics in a coherent and logical way. That is, Cerniglia et al (1990) indicates that sequential progression helps develop a topic, but too many new topics may distract the reader from the main idea. The following texts correspond to the finding.
(1)

He has a vigorous personality. (2) He can bring many laughs to our classroom. (3) He is a pistachio nut. (4) In the class, he can let the boring lesson to be more interesting. (5) In the sad time, he can comfort the bad mood friend to make his sadness go away. (6) He is my best friend, John. (7) One time, there is an important exam to us. (8) Everyone prepare for a long time toward that they can get the high score. (9) Fastly, the day is coming. (10) When I have confidence to do the test, the bad thing happened. (11) I got a bad sick so that I cant go to school to do the test. (12) Suddenly, all efforts are going for nothing. (13) Im very sad to miss the exam. (14) But John said We should look on the bright side. (15) There are many tests as important as this time, I dont need to care so much this time. (16) Gradually, I calm down my mind and thank to John for his care. (17) Its a good thing to have a best friend. (Sample L7) 1 2 3 4 5 6 7 8 9 1. he 1 2. he 1
39

3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17.

he he he he an important exam everyone the day the bad thing I all efforts I John many tests I to have a best friend

1 1 1 1 2 3 4 5 6 7 6 1 8 6 9

It is apparently to tell from the plotted table that the topic had been shifted away from its original topic. 8 sequential progressions occurred from sentence7 to sentence17. In spite of the high occurrence of sequential progression, each sequential topic was not directly related. For instance, topical subject an important exam in sentence 7 barely had relevance with everyone in the next sentence. Furthermore, the author described an event in second paragraph, yet, too many information was brought about which finally led to a digression of the topic focus. For example, sentence 7 to sentence 13, which emphasized on the important exam and the author himself, had nothing relating to the original topic of the essay. Many low-rated essays displayed such digression problems caused by employing too many sequential progressions. Take the following sample, for instance:
(1)

Tom, my best friend, who distribute himself to light my life up. (2) A coward came up to my mind when first time I met him, . (3) His skin is like snow white. (4) Once he carried pile of books, it seems that he would break down anytime. (5) I came to help and said Do you need help. (6) He said Yes, thank you. (7) From then on, we started to chat. (8) Gradually, I found he was not the person I thought before. (9) He help each other whether he likes, know about literature and good at sports.
40

(10)

One day, when I went home, three men holding knife try to kill me. (11) In that moment, I was so scared, finding my feet were so deep that I cant move. (12) Suddenly, a sound Ive ever heard said Here comes the police! (13) That three men run away instantly. (14) I luckily got saved. (15) he said then When best friend face to difficult time, I will save him to all my strength. (16) What he said touch my heart so deeply. (17) If best friend can be defined, Toms spirit for friend will be the one. (Sample L14) 1 2 3 4 5 6 7 8 9 1. Tom 1 2. coward 2 3. His 3 skin 4. he 1 5. I 4 6. he 1 7. we 5 8. he 1 9. He 1 10. three 6 men 11. I 4 12. a 7 sound 13. three 6 men 14. I 4 15. he 1 16. what 8 he said 17. his 9 spirit 8 sequential progressions in total were found in this essay, but none of them was directly related sequential progression. The topic had been digressed into the event happened between the author and the three men in the second paragraph. Worst of all, the final topical subject had never traced back to the main theme. Similar to the digression problem, the following sample extracted from low-rated essays contained tow topic sentences. Friends like a mirror who can light your bad habits and might change your life. (2) He or she may also help you when you are depressed. (3) Without question, friends are important in our life. (4) But your friend may not be a person, it can be a book, your pet or anything surrounding by you. (5) Example for me, I feel my best friend is comics. (6) Although they cant talk to me or give me a hug when I am disappointed,
41
(1)

they can cheer me up and make me happy. (7) In different to my friends in school, as soon as I read comics that I will put many unhappy things behind in my mind. (8) Although I feel comics are my best friend, my best friends are not only comics but also my friends in school. (9) My friends in school can teach me schoolwork or keep me company when I am sad. (10) But the important thing is that I can pour my heart to them. (11) In my life ups and downs, I can count on them. (12) I am happy that I have many friends. (Sample L20) 1 2 3 4 5 6 1. friends 1 2. he or 1 she 3. friends 1 4. your 2 friends 5. my best 3 friends (comics) 6. they 3 7. I 4 8. my best 5 friends (comics and school friends ) 9. friends 6 in school 10. I 4 11. I 4 12. I 4 By closely examining the sample, it is clear to find out that two topic sentences were attributed in sentence 5 and sentence 8 respectively. In the first paragraph, the writer indicated that she think her best friend is comics. However, in the second paragraph, she indicated that her best friends also include friends in school. At the concluding sentence, it was ended up with many friends. The focus of the essay had been shifted from my best friends to many friends. In sum, it is noteworthy that the digression occupied a great percentage in low-rated essays. The problem of incoherent and illogical problems found in low-rated essays is probably due to too many indirectly related or unrelated sequential progressions. Detailed discussion of extended parallel progression
42

The frequency of extended parallel progressions showed no big difference between high-rated essays and in low-rated essays. Nevertheless, several differences such as the extended parallel progressions that were placed in different paragraph has been found between high-rated and low-rated essays. For example, extended parallel progressions were often found in the conclusion part in high-rated essays. That is, high-rated essays tend to put extended parallel at the end of the paragraph in order to restate the main idea or to recall the readers memory about the main idea of the whole essay. However, low-rated essays failed to connect the idea through the whole essay by utilizing extended parallel progression. Some low-rated essays even contained no extended parallel progression. The following texts are extracted from low-rated essays that employed no extended parallel progression. Essay without extended parallel progression (1) When we are living in the world, we have friends. (2) They can play with us or share something with us. (3) A proverb say A near friend better than a fare dwelling kinsman. (4) Its meaning to make a good friend is important. (5) I have many friends, but know my heart is very little. (6) My best friend is Tony, who is a nice guy. (7) When I want to do something bad, he will advise me to think before I act. (8) And when my mood is bad, he will care about me. (9) Because my grade is bad, so he often asks me to study with him. (10) Recall to that all things, he is really my best friend because he cares about me and often do something good for me. (11) And we communicate now. (12) When we each other have hard time, we will help each other. (Sample L2) 1 2 3 4 5 6 7 1. we 1 2. they 2 3. a 3 proverb 4. its 4 meaning 5. I 5 6. my 6 best friend 7. he 6 8. he 6 9. he 6 10. he 6 11. We 7 (Tony and I) 12. we 7
43

Two essays in total, which belong to low-rated essays, conducted no extended parallel progression. One of the two essays is illustrated above, which only employed parallel and sequential progressions. It is apparently that the theme of the essay had been distracted from the original theme. That is, the topic in sentence 10 (he) was changed into we in sentence 11 and 12. And it had never returned to the original topic. In contrast to low-rated essays, high-rated essays tend to employ extended parallel in the conclusion paragraph. The following examples are extracted from high-rated essays. Extended parallel progression in concluding sentence
(1)

People say that A best friend is a color that brightens up your life. (2) Undoubtedly, when you get into trouble or sometimes in a bad mood, your best friend will always stay with you. (3) My best friend, Young, showed up when I was 13. (4) She was my new classmate in junior high school first grade, sitting next to me. (5) Just the opposite to me, Young is a gentle and quiet girl, always with smile on her face. (6) Suddenly, we became the greatest friend. (7) We seldom quarrel because we both realized the personality of each other and know how important the apologizes. (8) But until final year of junior high school, a girl who was disliked by our class joined our friendship. (9) She told something bad about me to Young and also told something terrible about Young to me, too. (10) So we quarrel and didnt talk to each other until we graduate. (11) There after month, one day I got a phone call from Young, she told me that she was misunderstanding me. (12) She wanted to said sorry to me. (13) I was so surprise that I even cant reply anything to her. (14) After she explained all the thing about the girl to me, I was so happy about we can be best friend again. (15) Young, in Chinese means the sun. (16) Until now, thought we dont study at the same school, we are still the friend which will always supports each other. (17) Young is the sun of my school life, brightening it up. (Sample H3) 1 2 3 4 5 6 7. A best 1 friend 8. Your 2 best friend 9. My best 3 friend 10. She 3 11. Young 3
44

12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23.

We We A girl She We She She I We Young We Young

4 4 5 5 4 3 3 6 4 3 4 3

This sample contained 5 extended parallel progressions. The first progression occurred between sentence 5 and 10. The second progression lay between sentence 7 and 10. The third one could be found between sentence 10 and 14, while the fourth progression was between sentence 12 and 15. The last progression was located between sentence 15 and 17. What is worth discussing is that not only how many parallel progressions were used frequently in this essay, but how they were used. High-rated essay tend to utilize extended parallel in the concluding sentence. It is a way to trace back to the main topic. Moreover, it was found that the concluding sentence has high relevance to the opening sentence. As shown in the Sample H3. The concluding sentence (Young is the sun of my school life, brightening it up) is corresponding to the first sentence (People say that A best friend is a color that brightens up your life.). Similar example was also found in the following sample extracted from high-rated essays.
(1)

Jenny, me playmate since childhood, is my best friend. (2) I first met her when I was 6 years old. (3) She and her family immigrated to Taiwan from Brazil. (4) Though she and I looked so different and spoke different language, we became friends almost as soon as we saw each other. (5) We found out that she loved insects, just like me. (6) Since then, we often went to parks to catch butterflies and grasshoppers, and kept caterpillars as pets. (7) Though we can hardly understand each others language, we communicate with each other with body languages and laughter. (8) Jenny moved away when I was 11 years old. (9) But we still wrote letters to each others. (10) In her letters, she talked about the new things she met in Canada and I told her the things happened in Taiwan. (11) She sometimes send me photos and Canadian salmon and Brazilian coffee beans. (12) We often talks about our future. (13) As time went by, I found she and I have both become mellower and mellower. (14) No matter how long were apart, Jenny will always be my best friend forever. (Sample H8) 1 2 3 4
45

1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14.

Jenny I she and her family we we we we Jenny we she she we she and I Jenny

1 2 3 4 4 4 4 1 4 1 1 4 4 1

Three extended parallel progression could be found in the above text. It is notable that extended parallel topics used here focused on the theme this writer was talking about (Jenny). Meanwhile, a extended parallel progression was used in the concluding sentence ( Jenny will always be my best friend forever), which was corresponding to the first sentence ( Jenny, me playmate since childhood, is my best friend.) On the contrary, unrelated concluding topics were often found in low-rated essays.
(1)

My best friend is a person that very love singing. (2) His name is Nana. (3) Most people will think that it must be a girls name but he is a man. (4) Why he and I will be best friends? (5) Because we all love singing. (6) I often sing a song with him in our classroom. (7) And we would sing louder and louder until everybody cant endure us. (8) Our classmates got headache about us but we found pleasure in it. (9) We not only singing together but sharing the new information of new songs. (10) In fact, his voice is much better than me. (11) But we didnt care about it. (12) Because when you sing with your truly feeling, the song will still pleasant to hear. (Sample L8) 1 2 3 4 5 6 7 1. my 1 best friend 2. his 2 name 3. he 3 4. he and 4 I 5. we 4 6. I 5 7. we 4
46

8. 9. 10. 11. 12.

we we his voice we The song

4 4 6 4 7

Two extended parallel progressions occurred in the above sample. However, the topic at the end of the essay had been drifted away from the original topic. we was the focus the writer tried to emphasize here, but the last sentence failed to draw readers attention back to the original topic we. The following sample is another example that contained an unrelated concluding sentence.
(1)

Friends are the most valuable present in everyones lifetime which full of depression. (2) My best friend is Eddie, whose behavior is well. (3) He always encourage me when I am sad. (4) I tell him my difficulties which bother me and depressed me every time. (5) He is endured with the natural talent for talking something to comfort my emotion. (6) There is no other person I deemed to be best friend prior to knowing him. (7) Eddie is a kind-hearted man who dedicates himself to the children in countryside. (8) He applied what he had learned to every kid patiently. (9) He also shared his feeling with me. (10) Gradually, I was influenced by him. (11) Now we are steep in helping people. (Sample L18) 1 2 3 4 5 1. friends 1 2. my best 2 friend 3. he 2 4. I 3 5. he 2 6. no other 4 7. Eddie 2 8. he 2 9. he 2 10. I 3 11. we 5 Only two extended parallel progressions were identified respectively between sentence 3 and 5, sentence 5 and 7. It is interesting to note that the last sentence (Now we are steep in helping people) had little reference to the first sentence (Friends are the most valuable present in everyones lifetime which full of depression) or to the
47

topic sentence (My best friend is Eddie, whose behavior is well). Unrelated concluding sentence often leave an impression that the essays has not been finished yet. Tale Sample L18 for example, the last two sentences indicated the fact that the writer was influenced by his best friend and want to help people in the future. Yet, little information was given to connect his best friend with the later theme mentioned in the last sentence (Now we are steep in helping people). That is, it seems that the topic could be further developed since the theme has not been traced back to the original topic my best friend. The relevance of sequential topics has effect on the use of extended parallel progression There is another interesting comparison found between high-rated essays and lowrated essays. That is, the level of relevance of sequential topics which lie within extended parallel progression affects the construction of coherence. According to the definition given by Schineider and Connor (1991), extended parallel progression refers to any sentence topic that is interrupted by at least one sequential topic before it returns to a previous sentence topic. In addition, as discussed in previous sequential progression section, sequential topics are classified into directly related, indirectly related and unrelated sequential topics. The researcher found that low-rated essays employed more unrelated sequential topics before it returned to the original topic. While high-rated essays contained more directly and indirectly related sequential topics before it returned to its original topic. The following samples contained the most extended parallel progressions among high-rated essays. Not only did these two essays put extended parallel progression in concluding sentence, but each extended parallel progression contained directly and indirectly related sequential topics before returning to the previous topics.
(1)

I used to be alone when I was in junior high school. (2) I was shy and quiet then so I didnt have any friend. (3) Until one day, she came to me. (4) Her name is Grace. (5) She is a very outgoing and energetic girl. (6) Our first conversation is unforgettable. (7) She said hi, my name is Grace. Since this moment you are my friend. (8) I burst out a laughter as soon as she finished what was going to say. (9) She laughed at me, too. (10) Since then, no matter where we went or what we did, we always be together. (11) We shared out little secrets to each other. (12) If we had something unhappy, we would cheered us up. (13) Time passed so quickly, three years in junior high were about to finish. (14) Both of us were very sad to say goodbye. (15) We will always be in touch. (16) Our friendship would last forever. (17) Grace will be my best friend forever. (Sample H10)
48

1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17.

1 I I

2 she

7 1 1 2 3

her name she our first conversation she I she we we we time both of us we Grace

2 4 2 1 2 5 5 5 6 5 5 Our 7 friendship 2

Total 5 extended parallel progressions are found here. They were distributed between sentence2 and 8, sentence3 and 5, sentence 5 and 7, sentence 9 and 17, sentence 12 and 14. As discussed in previous section, sequential progression can be classified into directly related and indirectly related sequential topics. In this Sample H10, the parallel progression occurred between sentence 2 and 8 contained three sequential topics (she, her name, our first conversation). she in sentence 3 was indirectly relate to the topic her name in sentence 4. Besides, three sequential topics (we, time, our friendship) occurred between sentence 9 and 17 were also related. For example, the sequential topic we in sentence 15 was also indirectly related to our friendship in sentence16. The following Sample H11 is also extracted from high-rated essays employed 4 extended parallel progressions, which occur between sentence 2 and 4, sentence 10 and 17, sentence 12 and 15, sentence 13 and 16. Three sequential topics (something, I and we) were found between sentence 10 and 17. Many directly related topics could be found. For instance, the theme I and rheme Puki in sentence 13 became the theme we in sentence14. Meanwhile, the theme horrible scene in sentence 15 has relevance to the theme a car fast hit Puki in sentence 14. As discussed in sequential progression, sentences 13 to 15 were topic-comment relation, which is regarded as directly related sequential progression.

49

(1)

It was a sunny day which I ran into her. (2) She had beautiful and round eyes. (3) What attracted me was that her particular temper. (4) She was like a peacock, glance and proud. (5) I nicknamed her Puki. (6) Little did I know that she will play such an important role in my childhood. (7) Magically, it seemed that she realized how I feel at any time despite that she was just a puppy. (8) When I was sorrowful, she sat down beside me and licked the tears which rolled down my cheeks. (9) On the contrary, when I was delightful, Puki also jumped happily with me. (10) I did everything with her, sharing my feelings with this lovely dog. (11) However, something changed all of a sudden. (12) A miserable event separated Puki and me. (13) One day, I walked Puki as every day. (14) We went through the road but a car fast hit Puki. (15) Now, the horrible scene would appear vividly when I think of Puki. (16) Since the day, I have lost her forever. (17) Although the event passed ten years ago, Puki is still my best friend in my memories forever. (Sample H11) 1 2 3 4 5 6 1. it 1 2. she 2 3. what 3 attracted me 4. she 2 5. her 2 6. she 2 7. she 2 8. she 2 9. Puki 2 10. her 2 11. something 4 12. miserable 4 event 13. I 5 14. we 6 15. horrible 4 scene 16. I 5 17. Puki 1 In sharp contrast, low-rated essays returned its topic to the original theme after many unrelated sequential progressions. The following texts are examples extracted from low-rated essays.
(1)

He has a vigorous personality. (2) He can bring many laughs to our classroom. (3) He is a pistachio nut. (4) In the class, he can let the boring lesson to be more interesting. (5)
50

In the sad time, he can comfort the bad mood friend to make his sadness go away. (6) He is my best friend, John. (7) One time, there is an important exam to us. (8) Everyone prepare for a long time toward that they can get the high score. (9) Fastly, the day is coming. (10) When I have confidence to do the test, the bad thing happened. (11) I got a bad sick so that I cant go to school to do the test. (12) Suddenly, all efforts are going for nothing. (13) Im very sad to miss the exam. (14) But John said We should look on the bright side. (15) There are many tests as important as this time, I dont need to care so much this time. (16) Gradually, I calm down my mind and thank to John for his care. (17) Its a good thing to have a best friend. (Sample L7) 1 2 3 4 5 6 7 8 9 18. he 1 19. he 1 20. he 1 21. he 1 22. he 1 23. he 1 24. an 2 important exam 25. everyone 3 26. the 4 day 27. the 5 bad thing 28. I 6 29. all 7 efforts 30. I 6 31. John 1 32. many 8 tests 33. I 6 34. to 9 have a best friend Only three extended parallel progression was included in this sample. Nevertheless, five sequential topics (An important exam, everyone, the day, the bad thing, I, all efforts) were found within sentence 6 and sentence 14. Most interesting of all, it was investigated that the essay that contained many unrelated sequential progressions within one extended parallel progression. The current sample L7 has been discussed
51

fully in the previous section in terms of unrelated sequential topics. A similar sample extracted from low-rated essays is discussed as follows.
(1)

Clicking my keyboard, I obtain abundant knowledge. (2) With computer, we can search a lot of information in different categories. (3) I cant imagine that what I can do without computer and Internet. (4) Thanks to the Internet, many people upload their research to the Internet. (5) By means of it, we can learn a lot of unknown information. (6) However, there is something to be desire that so many misleading knowledge on the Internet. (7) Someone who wants to know about the truth is misled by these rampant bullshit. (8) But tell the truth, the Internet is really gives a big hand to my study. (9) Needless to say, the Internet is my best friend. (10) Generally speaking, we are now living in a information-exploding generation. (11) How we catch up with the changing? (12) (You) just having good contact with the Internet. (13) My best friend will always give me good direction. (Sample L9) 1 2 3 4 5 6 7 1. I 1 2. we 2 3. I 1 4. Many 3 people 5. we 2 6. somethi 4 ng 7. someon 5 e 8. interne 6 t 9. interne 6 t 10. we 2 11. we 2 12. you 7 13. My 6 best friend 4 extended parallel progressions were distributed between sentence 1 and 3, sentence 2 and 5, sentence 5 and 10, sentence 9 and 13. Meanwhile, three sequential topics (Something, someone, internet) lay in the extended parallel progression. Besides, these three sequential topics were not directly related. Three sequential topics (Something, someone, internet) in Sample L9 showed little relevance to each other. Therefore, it was hard for readers to connect the ideas together. It is investigated that the difference in the use of extended parallel progression
52

between high-rated essays and low-rated essays lies in the relevance of sequential topics contained within extended parallel progressions. That is, the sequential topics high-rated essays used before returning to its original topic have high relevance to each other. Yet, low-rated essays contained more unrelated sequential topics before returning to its original topic. In other words, low relevance of sequential topics contained within extended parallel progression may digress the original topic. Even though extended parallel progressions are used to recall the main idea, it is hard for readers to connect all the irrelevant sequential topics to the original one.

Conclusion & Implications

This chapter presents the conclusion of the study. In 5.2, summary of the findings is reported in order to answer the research questions proposed in chapter one. Followed by the summary are pedagogical implications for EFL/ESL writing instructions. In addition, limitation of the study is discussed and some suggestions for future studies are provided as well. Summary of the Findings The purpose of the study is to investigate the rhetoric patterns and topical structures employed in students writings. This section summarizes the features found in the current study, which attempts to answer the following research questions. 1. What were the rhetoric patterns employed in composition produced by Taiwanese EFL senior high school students? 2. Is there any difference in terms of thematic progression between high-rated compositions and low-rated compositions? 3. How were the thematic progression proposed by Lautamatti (1987) carried out in the compositions? Rhetoric patterns found in students essays Five features of rhetoric patterns were found in students essays based on the placement of the thesis statement and the arrangement of the supporting ideas. The first type is the one which thesis statement appears in the beginning sentence, known as Thesis in the first sentence of the beginning paragraph (TFB). The second type refers to the thesis statement that is placed somewhere in the beginning paragraph. (TB). The third one which place thesis statement at the end of introduction paragraph is considered as thesis in the last sentence of the beginning paragraph (TLB). The forth rhetoric pattern refers to a postponed thesis statement in the second paragraph, which place thesis statement in the middle of the essays (TM). The last type refers to those writings with a delayed thesis statement in the end of the essays (TE). Notably, the first three types, including TB, TFB, and TLB are classified as deductive pattern,
53

which thesis statement is introduced in the early stage of the writing. TE, which contains a delayed thesis statement, is regarded as inductive rhetoric pattern. TM, which the thesis statement is placed in the middle of the writing, is regarded as nondeductive and non-inductive pattern. Examining the writing samples in the current study, deductive pattern occurred frequently in samples. That is, students preferred to introduce thesis statement in the early stage. Very few samples include the inductive pattern. In addition, the percentage of samples using TM (24%) was almost equal to that of samples using TB (25%). It was found that students tend to start writing with a brief opening statement. Proverbs and definitions were found majorly employed as an opening statement. TM developed the opening statement such as proverbs into a whole and well-developed paragraph. Therefore, the thesis statement was postponed till the second paragraph, with thesis statement appearing in the middle of the essay. Thematic progressions found in students writings In quantitative analysis, it was found that high-rated essays employed more parallel progressions than low-rated essays whereas low-rated essays employed more sequential progressions than high-rated essays. In addition, no difference in frequency of using extended parallel progression between high-and low-rated essays. In qualitative analysis, significant difference in usage of thematic progressions was found between high-rated essays and low-rated essays. In terms of parallel progression, high-rated essays were found to have competence in maintaining the focus of the essay by using consistent parallel topics. Moreover, instead of using pronouns, the name of the person being described in the essays was mentioned time to time in order to recall the readers mind. High-rated essays were found successfully leave a stronger impression of the essay to the reader while low-rated failed to maintain a consistent focus of the essay. Furthermore, more sequential parallel progressions were found in low-rated essays than in low-rated essays. Nevertheless, according to the definition proposed by Scheider and Connor (1991), sequential progressions are comprised directly related, indirectly related and unrelated sequential topics. More indirectly related and unrelated sequential topics were found in low-rated essays. Several directly related sequential topics were found in high-rated essays. The most common type of directly related sequential progressions was topic-comment relations. Topic-comment sequential topics were used to develop the main idea, and the topics were closely related. Yet, most sequential progressions found in low-rated essays were indirectly related or unrelated sequential topics, which digressed the focus away from the original topic. In addition, although extended parallel progressions were used equally frequent both in high- and low-rated essays, the arrangement and usage of extended parallel progressions were found different
54

between high-rated and low-rated essays. That is, in high-rated essays, extended parallel progressions were often placed at the end of the essay so as to ring a bell. Moreover, low-rated essay failed to recall the main theme of the essays at the end of the essay by using parallel progressions, the digression was pretty obvious. In conclusion, the problem of incoherence and digression were found in low-rated essays by utilizing topical structural analysis proposed by Lautimatti (1987). Pedagogical Implications The findings in the current study have some direct implication for EFL/ESL writing instructions. In Taiwan, it is common that English composition is taught based on sentence-level accuracy (Yao, 1989). The grammatical accuracy plays a significant role in the grading criterion of a good English writing. One suggestion of the study is that EFL teachers should pay attention to the distinctive feature of students writings. As presented in the findings, some students tend to start the writing with an introduction while an important thesis statement is left behind. Being aware of the rhetoric features of students writings, teachers will be able to help students be aware of the different conventions preferred by L1 and L2 students. For example, Chinese rhetoric values proverbs in wirings while English rhetoric regards proverbs as clich. By investigating the distinctive features of students writings, teachers can detect the preferred pattern employed by students so as to guide them with a proper instruction. Furthermore, Raimes (1999) proposed five principles to improve writing pedagogy concerning contrastive rhetoric studies. One of the principles is the recognition of students diversity. Namely, every student is different. Teachers need to realize that not all approach and procedures satisfy students. The emphasis on complexity of composition and diversity of students makes it imperative for researchers and educators not to seek a universal prescription of language teaching. Another principle is the recognition of learners process. Raimes(1999) noted that language teachers need to know about the process of how students learn a language and how writers construct a writing product. Since students approaches to writing in English are undoubtedly shaped by their educational backgrounds and rhetoric patterns (Spack, 1997, p.772), teachers should take learners process in learning and writing into account when applying to teaching. Therefore, teachers need to be aware of students writing process and products in order to design a writing approach that most satisfy the students. In addition, teachers are suggested to look writings beyond sentencelevel. Students writings should be emphasized more on the discourse level rather than on grammatical level. In the previous studies, an error-free writing is emphasized at the expense of coherence and logic. The finding of the present study suggested that the coherence of writings can be examined by utilizing Topical Structure Analysis
55

proposed by Lautimatti (1987). Previous studies have acknowledged TSA as a criterion of coherence and TSA has direct relation to the quality of writings. Many studies employed TSA as a tool of examining the coherence of writing. Additionally, TSA is used as a revision tool in students writing. Connor and Farmer (1990 ) employed TSA as a revision strategy for ESL writers and positively improved writing coherence by utilizing TSA. Similarly, Chiu (2004) in her case study help s student develop coherence based on TSA. Meanwhile, TSA is acknowledged as an effective way in improving writing coherence. In conclusion, TSA is not only an effective tool for teachers to investigate the coherence problems but an effective revision strategy for students to improve writing ability. As shown in the present study, it is investigated that improper use of parallel topics, unrelated sequential topics and inappropriate arrangement of extended parallel progressions can leads to digression and incoherence. Consequently, a mended writing instruction concerning to students problem can be provided. Moreover, TSA can also be taught as a reference for students to examine their writing quality as well as to improve their writing. Limitations of the Study There are some limitations in this study. One of them is that the sample collected was from one school only. The study failed to reflect the students diversity since data is collected from one single school. The number of students was rather small. Only 96 samples were recruited in investigating rhetoric pattern, and only 35 were selected in investigating thematic progression. In order to give a convincing and valid result, a larger number of samples should be collected. Second, only one rater is recruited in the current study. Since the analysis was quiet complex and sometimes required other proficient exports to identify and classify. It is suggested that two or more raters who have related background should be recruited in the analyzing process. Suggestions for Future Studies As mentioned before, the study lacked a large number of samples to provide a convincing result. More samples are suggested to be included in future studies. In addition, the samples were gathered from one single school in southern Taiwan. It is better for the future study to include diversity of students such as students from other places. Furthermore, more raters who have undergone similar analysis training should be included in the analyzing process. Moreover, the students learning background should be investigated as well. Since the rhetoric patterns are likely be influenced by L1 background, EFL students learning background should be taken seriously. It is suggested that a follow-up interview be carried out after the writing task since other
56

factors influencing the way they constructing the writings may be detected. Conclusion The purpose of the study is to examine the rhetoric pattern preferred by Taiwanese students and investigate how topical structures are employed in students descriptive writings. The study, meanwhile, intends to improve writing coherence and provide insights in writing instructions. Despite the limitation, the study expects to assist writing instructors to explore the rhetoric features and thematic progressions revealed in Taiwanese students writings. More future studies are suggested to carry out involving a larger number of samples, a more well-constructed analyzing process, and a full investigation of students learning backgrounds in order to achieve future insight to EFL writing instruction.

References
Almaden, D. (2006). An analysis of the topical structure of paragraphs written by Filipino students.The Asia-Pacific Education Research, 15 (1), 127-153. Cahill, D. (2003). The myth of the turn in contrastive rhetoric. Writing Communication, 20(2), 170-194. Cerniglia, C., Medsker, K., & Connor Ulla. (1990). Improving coherence by using computer-assisted instruction. In Ulla Connor & Ann M. Johns (Eds.), Coherence in Writing: Research and Pedogoical Perspectives (p. 187-207). Alexandria, Virginia: Teachers of English to Speakers of Other Languages, Inc. Chang, V. W. (1997). Freshman English composition: An error analysis from the discourse perspective. Taipei: Crane. Chen, S. L. (1994). Handbook for the elementary school students composition. Taipei: Quan Yuan. Chi, C. H. (2004). Analysis of text structures employed by EFL beginner writers. Unpublished master dissertation. National Yunlin University of Science and Technology. Taiwan. Chiu. Y. H. (2004). Coaching a student to develop coherence based on topical structure analysis: A case study. Journal of Language and Learning. 2(2), 154170. Chou, H. L. (1989). Contrastive rhetoric: Chinese and English. In S. M. Chang, D. S. Tseng, & B. C. Hwang (Eds.), Collection of Papers Presented in the Sixth Conference on English Teaching and Learning in the Republic of China. (pp. 251-269). Taipei, Taiwan: Crane. Connor, U. & M. Farmer (1990). Topical structure analysis as a revision strategy for ESL writers. In Barbara Kroll (Ed.), Second language writing: research insights for the classroom. Cambridge: CUP Hoenish, S. (2008). Topical structure analysis of accomplished English prose. Published masters thesis, The City University of New York, New York. Johns, A. M. (1986). Coherence and academic writing: Some definitions and suggestions for teaching. TESOL Quarterly, 20(2), 247-265. Kaplan, R. (1966). Cultural thought patterns in intercultural education. Language Learning, 16, 1-20. Kirkpatrick, A. (1997). Traditional Chinese text structures and their influence on the writing in Chinese and English of contemporary mainland Chinese students.
57

Journal of Second Language Writing, 6(3), 223-244. Lautamatti, L. (1987). Observations on the development of the topic of simplified discourse. In Ulla Connor and R.B. Kaplan (Eds.), Writing across languages, analysis of L2 texts, Reading, MA: Addison Wesley. Lee, I. (2002). Teaching coherence to ESL students: a classroom inquiry. Journal of Second Language Writing, 11, 125-159. Lin, S.O., Huang, T.S., Lin, C.I., & Chen, K.T. (1993). Construction and grading of English composition tests in the JCEE(). In Selected Papers from the Tenth Conference on English Teaching and Learning in the Republic of China. (pp. 557-609). Taipei: Crane. Liu, J. (2009). Intercultural rhetorical pattern differences in English argumentative writing. English as International Language Journal 4, 93-109. Makinen, K. (1992). Topical depth and writing quality in student EFL compositions. Scandinavian Journal of Educational Research, 36, 237-247. Matalene, C. (1985). Contrastive rhetoric: An American writing teacher in China. College English, 47 (8), 789-807. Nordquist, R. (1987). Passages: A Writer's Guide. New York: St. Martin's Press. Oshima, A., & Hogue, A. (1997). Introduction to academic writing (2ed.) London; Longman. Petric, B. (2005). Contrastive rhetoric in the writing classroom: A case study. English for Specific Purposes. 24, 213-228. Phuwichit, K. (2004). A topical structure analysis of essays written by fourth year English major students. Masters project, M.A. (English). Srinakharinwirot University, Bangkok, Thailand. Raimes, A. (1991). Out of the woods: Emerging traditions in the teaching of writing. TESOL Quarterly, 25(3), 407-430. Richards, J. et al. (1985). Longman dictionary of applied linguistics. London: Longman. Schneider, M., & U. Connor. (1991). Analyzing topical structure in ESL essays: Not all topics are equal. Studies in Second Language Acquisition 12, 411-427. Silva, T. (1992). L1 vs. L2 writing: ESL graduate students perceptions. TESL Canada Journal, 10(1), 27-48. Simpson, J. M. (2000). Topical structural analysis of academic paragraph in English and Spanish. Journal of Second Language Writing, 9(3), 293-309. Spack, R. (1997). The rhetoric construction of multilingual students. TESOL Quarterly, 31, 765-774. Spencer, Wang & Xing. (2008). Raising students' awareness of cross-cultural contrastive rhetoric in English writing via an e-learning course. Language Learning & Technology. 12(2), 71-93. Witte, S.P. (1983). Topical structure and writing quality: A study of the argumentative texts of college writers. Montreal, Canada. Wong, s. D. (1992). Contrastive rhetoric: An exploration of proverbial references in Chinese L1 and L2 writing. Journal of Intensive English Studies, 6, 71-90. Wu, S. Y. & Rubin, D. L. (2000). Evaluating the impact of collectivism and individualism on argumentative writing by Chinese and North American college students. Research in the Teaching of English. 35(2), 148-178. Yang, L., & Cahill, D. (2008). The rhetorical organization of Chinese and American students expository essays: A contrastive rhetoric study. International Journal of English Studies. 8(2), 113-132. Yao, L. C. (1989). An overview of instructional theory and research and its applications to EFL instructions in Taiwan. Taichung, Taiwan: Chung Hsing
58

University. Young, L. W. L. (1994). Crosstalk and culture in Sino-American communication. New York; Cambridge University Press.

59