(LUẬN văn THẠC sĩ) a topical structure analysis to create coherence in english essays of fourth year students at english department at vietnam national university, university of language and international studies

Table 1: Essays‟ score description 18 Table 2: The mean, mode, median of the numbers of T-units, paragraphs, and topical depths of low-rated and rated essays high-24 Table 3: The variati

FACULTY OF POST-GRADUATE STUDIES

NGUYỄN THỊ HỒNG

A TOPICAL STRUCTURE ANALYSIS TO CREATE

COHERENCE IN ENGLISH ESSAYS OF FOURTH YEAR STUDENTS AT ENGLISH DEPARTMENT AT VIETNAM NATIONAL UNIVERSITY, UNIVERSITY OF LANGUAGES AND

INTERNATIONAL STUDIES

(Phân tích cấu trúc chủ đề để tạo liên kết mạch lạc trong các bài luận tiếng Anh của sinh viên năm 4 khoa Anh, trường đại học

Ngoại Ngữ, đại học Quốc Gia Hà Nội)

M.A MINOR PROGRAMME THESIS

Field: English Linguistics

Code: 60220201

FACULTY OF POST-GRADUATE STUDIES

NGUYỄN THỊ HỒNG

A TOPICAL STRUCTURE ANALYSIS TO CREATE

COHERENCE IN ENGLISH ESSAYS OF FOURTH YEAR STUDENTS AT ENGLISH DEPARTMENT AT VIETNAM NATIONAL UNIVERSITY, UNIVERSITY OF LANGUAGES AND

INTERNATIONAL STUDIES

(Phân tích cấu trúc chủ đề để tạo liên kết mạch lạc trong các bài luận tiếng Anh của sinh viên năm 4 khoa Anh, trường đại học

Ngoại Ngữ, đại học Quốc Gia Hà Nội)

M.A MINOR PROGRAMME THESIS

Field: English Linguistics

Code: 60220201 Supervisor : Professor Hoàng Văn Vân

HANOI- 2016

entitled

A TOPICAL STRUCTURE ANALYSIS TO CREATE COHERENCE

IN ENGLISH ESSAYS OF FOURTH YEAR STUDENTS AT ENGLISH

UNIVERSITY OF LANGUAGES AND INTERNATIONAL STUDIES

submitted in partial fulfillment of the requirements for the degree of Master in English Linguistics Except where the reference is indicated, no other person‟s work has been used without due acknowledgement in the text of the thesis

Moreover, I hereby declare that all the essays which have been used as the primary source for the research has been allowed to be investigated by the students who wrote the essays and the teacher who rated the essays and all these essays haven‟t been used in any preceding research

First and foremost, I would like to express my sincere gratitude to Professor Hoàng Văn Vân, my supervisor, who has patiently and constantly supported me

through the stages of the study, and whose stimulating ideas, expertise, and suggestions have inspired me greatly through my growth as an academic researcher

Secondly, I would like to give my deep thank to Mr Lê Thành A (the

pseudonym name of the lecturer), the lecturer of Literature in English in Vietnam National University, University of Language and International Studies, and he is the rater of the essays which I used as primary source of data for my research

Last but not least, I am greatly indebted to my family, my friends for the sacrifice they have devoted to the fulfillment of this academic work

In the population of forty essays written by forth year students at ULIS, VNU, twenty-six essays including thirteen high-rated and thirteen low-rated ones were taken out for two objectives The first objective is to compare and contrast the physical features of two groups and the topical structures The second objective is to establish a possible relationship between the essay holistic evaluation and the type

of topical progressions used Data analysis has revealed the moderately physical differences between two groups in terms of t-units, topical depths, but no significant difference could be found in the numbers of paragraphs and topical structures In addition, findings also indicated that there was significant difference in the progression proportions used in two groups, especially in parallel progression in body paragraph and the subtypes of sequential progression in introductions and conclusions

ACKNOWLEDGEMENTS ii

ABSTRACT iii

TABLE OF CONTENTS iv

LIST OF TABLES AND FIGURES viii

INTRODUCTION 1

1 Rationale for the research 1

2 Aims and objectives of the research 2

3 Research questions 2

4 Scope of the research 3

5 Methods of the study 3

6 Significance of the research 4

7 Structural organization of the thesis 4

DEVELOPMENT 5

CHAPTER 1 - LITERATURE REVIEW 5

1.1 Theoretical background of coherence in written texts 5

1.1.1 Overview of coherence 5

1.1.2 Approaches to coherence 6

1.1.2.1 Cohesion 6

1.1.2.2 Genre 7

1.1.2.3 Semantic relation 7

1.1.3 Methods for examining coherence in a text 8

1.1.3.1 Latent semantic analysis 8

1.1.3.2 Topical structure analysis (TSA) 9

1.2 Theoretical background of TSA 9

1.2.1 Topical Structures 10

1.2.2 Topical progressions 11

1.4 Summary 15

CHAPTER 2 – RESEARCH METHODOLOGY 16

2.1 Research methodology 16

2.1.1 Restatement of research questions 16

2.1.2 Research approach 16

2.1.3 Principles/Criteria for intended data collection and data analysis 17

2.2 Research methods 17

2.2.1 Topical Structure Analysis method 17

2.2.2 Data collection techniques 18

2.2.3 Data analysis procedure 19

2.3 Summary 23

CHAPTER 3 - FINDINGS AND DISCUSSIONS 24

3.1 The physical features of essays 24

3.1.1 Findings on the essay physical features 24

3.1.1.1 Numbers of essay T-units, paragraphs and sub-topics 24

3.1.1.2 Findings on the topical structures of clauses 26

3.1.2 Discussion on the physical features of essays (Research question 1) 27

3.1.2.1 Numbers of t-units, paragraphs, and sub-topics 27

3.1.2.2 The topical structures of t-units in essays 28

3.2 Proportions of topical progressions in essays 28

3.2.1 Findings on the difference between low-rated and high-rated essays 29

3.2.1.1 Topical progressions used in all essay paragraphs 29

3.2.1.2 Topical progressions in introductions and conclusions 31

3.2.2 Discussion on the uses of topical progressions 33

3.2.2.1 Topical progressions in all essay paragraphs and body paragraphs 33

3.2.2.2 Topical progressions in introductions and conclusions 36

mood subject initial sentence element sequential progression(s) directly related sequential progression indirectly related sequential progression unrelated sequential progression

parallel progression(s) extended parallel progression(s)

Table 1: Essays‟ score description 18 Table 2: The mean, mode, median of the numbers of T-units,

paragraphs, and topical depths of low-rated and rated essays

high-24

Table 3: The variation of the frequency of T-units in

introductory, body and concluding paragraphs of essays

Figure 2: The difference on the portions of each type of topical

progression in low-rated and high-rated essays

29

Figure 3: The proportions of different types of progression in the

body paragraphs of two groups of essays

INTRODUCTION

1 Rationale for the research

Coherence seems to be a key indicator of the text quality as it deals with the global and local meanings and the deep organizational logic of the text The more coherent the texts are, the higher the text quality achieves In general, evaluating the text coherence is supposed as problematic because coherence has been considered

to be a subjective and hazy concept which is hard to learn and teach (Lee, 2002) As Halliday and Hassan (1976) indicated, the distinction between a text and a collection of unrelated sentence is a matter of degree; therefore, the holistic evaluation is quite subjective in measuring text quality because it lacks reliability and validity for not indicating particular strengths and weaknesses of learner texts However, subjective reader‟s evaluation on text is still a dominant method in

evaluating the text in teaching contexts

Nowadays, a number of investigations have been carried out to examine how coherence is achieved in learner texts and to help the teachers make sound judgments on student compositions One major method is to analyze the topic development of the high-rated and low-rated essays by applying the Topical Structure Analysis (TSA) developed by Lautamatii (1978) This method helps to examine the links between the reader‟s evaluation and the topical developments of the text to explicitly state the text features for classifying the text quality Overall, different patterns (topical progressions) of topical structure appear to provide good

predictors of student writing quality

In general, this research investigated the relationship between the holistic evaluation and the TSA of the student essays to find out the differences in topical structures and the topical progressions used between low-rated and high-rated

essays

2 Aims and objectives of the research

The research population is the collection of forty essays written by forth-year students at ULIS, VNU The research aims at investigating the relationship between holistic evaluation and the TSA of the essays by examining thirteen high-rated and thirteen low-rated essays for two main objectives The first objective focuses on the physical features of the essays including the numbers of t-units (the shortest unit consisting of one independent clause together with whatever dependent clauses are attached to it), paragraphs, topical depths as well as the topical structures in the essays The research examines the differences between low-rated and high-rated essays in terms of physical organization and the various combinations of the mood subjects (mood S), the topical subjects (topical S), and the initial sentence element (ISE) of the clauses in the essays in order to identify the contribution of these features to essay quality The second objective examines the relationship between the text quality and the topical progressions used in student essays It focuses on examining the proportions of topical progression types (sequential, parallel, and extended parallel) in two group essays Particularly, it separately investigates the proportions of progressions used in the introductions, the body paragraphs, and the conclusions of essays to have a view on the contribution of these progressions in the essays To sum up, the research aims to reveal the physical features, topical structures and progressions that contribute to the quality of the essays

3 Research questions

In order to achieve two main objectives, the research focuses on answering

two main following research questions:

Research question 1: How do the physical features in terms of lengths and

the topical structures reveal the differences between low-rated and high-rated essays

of Vietnamese students?

Research question 2: What is the relationship between holistic evaluation

and the topical progressions used in essays of Vietnamese students?

4 Scope of the research

The research focuses on investigating the English essays written by Vietnamese students Particularly the research explores the essays of the forth year students at ULIS, VNU written in their American Literature course in the year of

2013 These essays are chosen as the primary data for the research because of three reasons Firstly, these essays share many similarities including the same topic which

is about Buck‟s conflict in the story “The Call of the Wild” Secondly, these essays are commented and scored by the teacher, and sent back to the students through email, it brings two advantages for the researcher One is the convenience for collecting all the data without difficulty and the other is that the essays had been scored by the teacher; therefore they were an appropriate set of data for searching answers for the research questions

5 Methods of the study

The research applies the TSA framework as developed by Lautamatii (1978)

to identifying the relationship between clauses in the paragraphs The choice of topical structure as the focus of this study is motivated by the following considerations First, topical structure analysis probes an important aspect of texts, namely the patterns of maintenance and shifts of topics Such patterns contribute considerably to the text coherence, to the identification of what a particular stretch

of discourse is about, and, consequently, to the comprehensibility of texts Therefore, this is an objective method for analyzing the student writing Second, topical structure analysis allows for the quantification of data which makes the study more reliable, because the frequency and percentage of t-units, paragraphs, topical depths, topical structures and topical progression types in student essays are calculated based on the carefully analyzed of essays In summary, using TSA as a tool for analyzing student essays is the most suitable choice for this research

6 Significance of the research

In general, the research is highly significant for the teaching and learning of English writing in Vietnam The identification of physical features, topical structures and the dominant topical progression used in essays of Vietnamese students supplies understanding on the essay writing features of Vietnamese students In addition, comparing the topical progressions of highly rated essays and low rated essays has pedagogical implications for teaching students how to write a good essay Teachers can use topical structure analysis to help orient ESL students' writing to the underlying structures that build coherence

7 Structural organization of the thesis

The research consists of three main parts, part one: Introduction, part two: Development, and part three: Conclusion Part one briefly describes the rationale for conducting the research, two objectives of the research and the Topical structure analysis (TSA) method used Part two includes three chapters: Literature review, Research methodology, Findings and Discussions Chapter one focuses on the theoretical background as well as the review of previous research with the similar interest In the theoretical background section, the overview, approaches as well as the measurements of coherence are concerned In addition, it presents the background information of TSA In the review of related literature, there are the reviews of research on the proportions of topical progressions in student essays in different contexts and their results, and of the studies which examined the difference between low-rated and high-rated essays in terms of the use of topical progressions

In chapter 2, research method of the study is in some detail described It introduces the population and sample of the research, the data analysis procedure The third chapter, Findings and Discussions, focuses on answering two main research questions using the results from the data analysis And part three summarizes what has been found in the research, provides some implications as well as suggestions

for further research

DEVELOPMENT

CHAPTER 1 - LITERATURE REVIEW

1.1 Theoretical background of coherence in written texts

1.1.1 Overview of coherence

“There is little consensus on the matter of an overall definition of coherence”

as Grabe and Kaplan (1996: 67) indicated In general, the understanding of coherence develops through the time with the focus from text-based features such as cohesive devices (Halliday and Hassan, 1976) and semantic property of discourse (Van Dijk, 1977) to the reader-based features (Johnson, 1986)

In the traditional approach, linguists including Halliday and Hassan (1976), Van Dijk (1977) and Enkvist (1978) focus on the text-based features to explore text coherence On one hand, Halliday and Hassan (1976) focus on linguistic features of the texts According to them, coherence or texture is the feature of a text, which identifies the text as a total unit A text is considered coherent when the various parts of a paragraph are connected together by cohesive ties On the other hand, focusing on the semantic meaning of the text, Van Dijk (1977) and Enkvist (1978) identified coherence in a text by looking at the semantic relationship between sentences in the text Van Dijk (1977) indicated that coherence may be considered

as a “semantic property of discourses, based on the interpretation each individual sentence relative to the interpretation of other sentences” In addition, Enkvist (1990) also defines coherence as “the quality that makes a text conform to a consistent world picture and is therefore summarizable and interpretable”, and coherence is primarily related to the nature and property of the text

The understanding of coherence widens from text-based into reader-based features by Johnson (1986: 247) who defines coherence in written text is “a complex concept, involving a multitude of reader- and text-based features” According to Johns (1986: 250), coherence is “a feature internal to text, either in

terms of the linking of sentences (cohesion) or as the relationships among propositions in the text (sticking to the point)” Text-based coherence refers to an inherent feature of the text, which has been discussed in the preceding paragraph Reader-based coherence, in contrast, requires successful interaction between the reader and the text Following Johnson‟s point of view, many linguists also defines coherence in terms of reader-based features

In summary, coherence includes two main sides: text-based and reader based features These sides co-exist and create the writing coherence

1.1.2 Approaches to coherence

Coherence is an abstract concept It is, therefore, defined differently by different linguists In this section, three aspects of coherence including cohesion, genres, and semantic relation are focused to enlighten the scopes in which coherence has been studied

1.1.2.1 Cohesion

According to Halliday and Hasan (1976), cohesion is the major characteristic

of coherence Cohesion occurs where the interpretation of some elements in the discourse is dependent on that of another (Halliday&Hasan, 1976) Cohesion depicts how meaning-based relationship is set up by lexical and syntactic features These explicit lexical and syntactic features are known as cohesive devices, signaling the relationship in sentences and paragraphs Halliday and Hasan (1976) introduced five different types of cohesive devices in order to provide a guideline for studying and judging the cohesion and coherence of writing: reference, substitution, ellipsis, conjunction and lexical cohesion They contended that through analyzing the use of cohesive devices, one could evaluate or assess writing quality from the perspective of coherence

However, the cohesive theory proposed by Halliday and Hasan (1976) was challenged by Carrell (1982) and Johns (1986) who argued for the importance of

readers‟ background knowledge In other words, both the structure and content of the text and the readers‟ operation on the text should be taken into consideration This leads to the second approach to coherence which is genre approach

“literatedness” which Kaplan (1966) has considered as an approach by indirection

On the other hand, Farrel (1997a, 1997b) argues that school literacy in most English countries is a highly specialized discourse that “giving directly to the point” or being linear in its development Therefore, the relevant patterns in Vietnamese writing may be seen as digressive by some English readers When conducting an English essay, Vietnamese students need to pay attention to the rhetorical patterns

of English language However, as far as an essay organization is concerned, both English and Vietnamese essays share these three components: introduction, body and conclusion What make the difference are the things included in each part The current study approach to analyze the essay components separately to have deep view on the ways ideas organized in each part

1.1.2.3 Semantic relation

In identifying the coherence in writing, most researchers focus on the semantic relations in the text For example, Beaugrande and Dressler (1981:85) claimed that

relevance within the configuration of concepts and relations” Without the continuity, any piece of writing is just plain writing, without making much sense to the reader about the points it makes In addition, Celce-Murcia and Olshtain (2000) stated that the overall coherence of a longer text depends on the coherence within each paragraph or section of the text This textual relationship is partially a result of coherent organization of the propositions and ideas presented in writing

In particular, the study of „topic‟ in discourse becomes a promising branch of research in examining the semantic relation of essays Prague School linguists such

as Danes (1974), Firbas, (1974), and Mathesius (1975) have contributed to the analysis of extended texts, emphasizing the role of sentences in the context of whole texts, rather than any individual units of discourse Following the Prague School, Lautamatii (1987) developed TSA to examine the topic development in the discourse Obviously, there are many methods for identifying the semantic relation between sentences in a text However, two most promising methods are Latent Semantic Analysis (LSA) and TSA In the following part, brief reviews on these two methods are carried out, and the reasons for choosing TSA as the primary method in this research are stated

1.1.3 Methods for examining coherence in a text

Researchers have proposed various techniques by which teachers can measure coherence Most of these techniques are based either on statistical analysis, such as the LSA proposed by Foltz, Kintsch, and Landauer, (1998), and Lautamatti‟sTSA

1.1.3.1 Latent semantic analysis

LSA is a method for measuring textual coherence LSA is a fully automatic statistical technique for extracting and inferring relations of expected contextual usage of words in passages of discourse Word, sentence and passage meeting representation derived by LSA have been found capable of simulating a variety of human cognitive phenomena, ranging from developmental acquisition of

recognition vocabulary to word-categorization, sentence-word semantic priming, and judgments of essay quality and discourse comprehension (Landauer, 1998) The method is proved to be automatic and fast, permitting quick measurements of the semantic similarity between pieces of textual information However, this method is quite complicated and it requires much information about computers, the application of this method for teaching practice is not favored This technique is unlikely to be employed by teachers or students because it is time-consuming and require a considerable amount of training

1.1.3.2 Topical structure analysis (TSA)

Besides LSA, drawing on Prague School research, Lautamatti (1987) developed what she called TSA as a tool for analyzing the coherence of writing The use of TSA as a 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 In general, the study of TSA has gained interest among researchers in their desire to determine thematic development in paragraphs and styles of writing of individuals The detail information about TSA will be stated in the next part, which covers all the issues related Because it is the method which is applied to identify the relationship between holistic evaluation on the quality of the essays and the topical progressions used

1.2 Theoretical background of TSA

TSA refers to the analyzing of coherence derived by examining the internal topical structure of the paragraph as reflected by the repetition of key words and phrases In TSA, two important terms are topical structures and topical progressions

1.2.1 Topical structures

Topical structures are the combinations of the mood subject (mood S), topical subject (topical S), and initial sentence element (ISE) of the t-units in the discourse Firstly, mood S is the structural subject of the clause Its position is before the main verb or auxiliary verb in the clause And topical S is the lexical mood S which is directly related to the discourse topic In structural dummies such as there-sentence

or empty-it, the topical subjects are the noun phrases at the predicate, and these structural dummies are non-topical subjects In addition, non-topical subjects are also the lexical mood S which are not directly related to the discourse topic Non-topical linguistic materials include discourse connectives, illocution markers, modality markers, attitude markers, metatextual and commentary markers Discourse connectives are linguistic items indicating order and logical connectors

like consequently, however…Illocution markers are used to make explicit the illocutionary force of the statement concerned such as for example, to illustrate the point Modality markers illustrate the truth value of the information discussed including it seems probably, obviously, somebody suggests, I doubt whether Attitude markers such as I would like to, it seems futile to… are used to make explicit the writer own attitude Metatextual and commentary markers such as next,

we shall discuss, in later chapter, we will attempt to, are used to comment on the

discourse itself, language of the text or the organization of the text Finally, ISE refers to the initially placed discourse material in sentence, whatever its form or type It may be the topical S, mood S or non-topical subject

In general, there are five possible combinations of the mood subject, topical subject, and the ISE in a clause, and they are abbreviated as T1, T2, T3, T4, and T5

Type 1 (T1): ISE = mood S = topical S

Type 2 (T2): ISE ≠ mood S = topical S

Type 3 (T3): ISE = mood S ≠ topical S

Type 4 (T4): ISE = topical S ≠ mood S Type 5 (T5): ISE ≠ mood S ≠ topical S

The clause belongs to T1 type if the ISE, mood S, and topical S coincide One

example is the t-unit “The call of the wild (T1) was the best seller book which

brought fame to Jack London in 1903” In this t-unit, the noun phrase “The call of the wild” is the ISE, the mood S as well as the topical S of the clause, so the noun phrase “The call of the wild” is bolded and marked (T1) after it The “T2” type is

put after the topical S if the ISE is separated from mood and topical S, which

coincide The t-unit “In order to live in a wild life, Buck (T2) has an external

conflict with Spitz- another sled-dog during their trip to go to Northland” is an example The ISE of the t-unit is the adverbial phrase “In order to live in a wild life”, and the noun phrase “Buck” is both mood S and topical S of the clause The

topical structure of the t-unit is marked T3 if ISE and mood S coincide while topical

S separate There are two subtypes in type 3; in one variant, the mood S of the initial main clause is a dummy, with the noun phrase in the predicate representing

the topical S, as in the t-unit “it is the moment (T3) that Buck aware of the law of

wild life” The other variant, the initial main clause is non-topical subject, while the topical S appears in a following sub-clause For example, it is the t-unit “It seems

that Thornton’s Death (T3) releases Buck from the fragment bond with

civilization” in one essay in the sample The topical S belongs to T4 if ISE and

topical S coincide, while mood S is separate T4 often occurs in the t-unit which consist of the topical S in an initial sub-clause and a cohesive anaphoric pronoun as

the mood S in the main clause, the clause “When Buck becomes a sled-dog with a

struggle for survival, he (T4) must be used to the law of club and fang day by day,

lesson after lesson” is an example The last type, T5 occurs in the clause in which the ISE, mood S, and topical S are all separate For example, the clause “In

addition, it is also the moment (T5) that Buck recognizes the wildness, cold, and

ruthless in Spitz’s face”, the connective “in addition” is the ISE, empty-it is the mood S, and “the moment” is the topical S In summary, this part has made a brief

explanation on the topical structures and its types

1.2.2 Topical progressions

In general, discourse topic is the main idea discussed in the whole essay or text Subtopics are the subordinate ideas to support for the discourse topic, and they are directly related to the discourse topic The hierarchically arrangement of subtopics expresses the development of ideas to contribute to the discourse topic According to Lautamatii (1978), the way the written sentences in discourse relate to the discourse topic and its subtopics is called topical development of discourse Topical progressions which show the topical development is the sequence of developing one sub-topic, adding new information about it in the predicate of that subtopic, and then proceeding to develop another sub-topic in order to contribute to the development of the discourse topic In general, Lautimatti (Ibid.) proposed three types of topical progressions including parallel (PP), sequential (SP), and extended parallel (EPP) PP consists of two consecutive clauses with the same topical subject The use of PP is perceived by the readers as repetition of the same topics in the essay and reinforcement of the ideas, but, if used over-extensively and at the expense of EPP, may produce the impression of redundancy The second type is SP,

in which the topic of the next sentence is the comment of the previous sentence, or a new different topic And, the use of SP is perceived by the readers as development

of individual topic by adding details to an idea If used sensibly and in good balance with other types of progression, it may add to an overall good quality, whereas too much development for a sentence topic may distract the readers from the main ideas Thirdly, the topic in EPP is the same as the topic of the previous sentence but

is interrupted by some SP The use of EPP is perceived by the readers as good focus and explicitness of expression Absence or sparse use of EPP may produce an overall impression of an unfocused text which lack global meaning

In this research, three types of progressions in the paragraphs of the essays were identified using the TSA method In addition, based on research of Schneider

and Connor (1991), three subtypes of SPs also examined for further research on the differences between low-rated and high-rated essays

1.3 Review of the related studies

In general, TSA is considered as an objective method to check coherence and proportions of topical progression in students‟ texts (Gao,2012; Almaden, 2006; Barabas&Jumao-a, 2009; Carreon, 2006; Phuwichit (2004), Simpson, 2000; Lautamatti, 1987; Witte &Faigley, 1981) In this part, the research on how TSA identified the proportions of topical progressions in student texts as well as the use

of TSA as a tool for classification between low-rated and high-rated essays would

be reviewed

1.3.1 The findings of prominent studies

This section briefly presents the findings of previous studies that investigated the relationship between the progressions of topical structure and the assessors' judgments of essay‟s quality The two most prominent studies are the studies of Schneider and Connor (1991) and Witte (1983a, 1983b) In general, both of the research indicated that there are significant differences between the low-rated and high-rated essays in the frequency of the three topical progressions However, the findings of two studies on the progressions which distinguish between low-rated and high-rated essays are contradictory While Witte found that low-rated essays contained more SPs than high-rated essays, Schneider and Connor (Ibid.) found that highly rated essays contained a high proportion of coherence-building SPs and an EPP that helps return the essay to its main theme

According to Schneider and Connor (Ibid.), the differences in criteria for coding progressions led to the differences in the findings in the studies of Schneider and Connor (1991) and Witte (1983a, 1983b) And Schneider and Connor (1991:422) believed that "Not all sequential topics contribute equally to the coherence of a text" It is clear that “directly related” SP which the topic of a

sentence is related to the comment of the previous sentence and indirectly related

SP which the topic of the new sentence is partly related to the comment of the previous sentence (indirectly related topics) build coherence in the essays And the subtopics which are not directly related to the discourse topic (unrelated SPs) are not likely to contribute to create coherence in the paragraphs Therefore, deep investigations to the SP by examining three subtypes of SP contribute to the coherence of the essays especially in introductory and concluding parts

1.3.2 Findings of other research

Besides the research of Schneider and Connor (1991) and Witte (1983a, 1983b), there are other research which focused on searching the differences between low-rated and high-rated writings such as Connor and Farmer (1990), Makinen (1992), Burneikaite&Zabiliute (2003)Liangprayoon, Chaya&Thep-ackraphong (2013)

As Connor and Farmer (1990) indicated, if a writer logically makes use of SP and make good balance with the other types of topical progression, it may add to the overall good quality Additionally, essays are perceived as better quality and particularly more coherent when the use of the PP is in good balance with the use of EPP (Burneikaite&Zabiliute, 2003) And the texts are considered as poor quality essays when PP is over-extensive and EPP is insufficient

Moreover, by applying TSA to short compositions written in English, Makinen (1992) concluded that good writers have the ability to develop the topics

in their compositions more evenly across several topic levels than mid-quality writers and especially the 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 In addition, the research on the effectiveness of TSA instruction of Liangprayoon, Chaya&Thep-ackraphong (2013) found that the proportions of topical progressions used in the highly-rated essays were not different from those used in the low-rated ones

Further, both successful and less successful students employed the SP the most in

the essays

This chapter focuses on theoretical background and literature review of the

research The theoretical background discusses the term coherence and the TSA

framework to approach to analyze the texts On one hand, text coherence is

expressed through text quality based on reader‟s evaluation On the other hand,

TSA includes the analysis of topical structures and topical progressions Therefore,

this research examines the relationship between reader‟s evaluation and TSA of the

essays Through the literature review, there are many differences in the level of

topical depths, the proportions of topical progressions, and the balance between

topical progressions to the level of coherence in an essay, so this research aims to

identify the differences between low-rated and high-rated essays by using TSA

CHAPTER 2 – RESEARCH METHODOLOGY

2.1 Research methodology

2.1.1 Restatement of research questions

The two research questions raised on pages are repeated here for exploration:

Research question 1: How do the physical features in terms of lengths and the topical structures reveal the differences between low-rated and high-rated essays of Vietnamese students?

Research question 2: What is the relationship between holistic evaluation and the topical progressions used in essays of Vietnamese students?

on the statistics to investigate the differences between low-rated and high-rated essays

2.1.3 Principles/Criteria for intended data collection and data analysis

The research population is the collection of 40 essays from two classes E11QH2009 and E12QH2009 in ULIS, VNU These essays were written in their 8thsemester in their Literature in English course in the year of 2013

These essays are taken as the primary data for the research for two reasons Firstly, all these essays are written to complete the task of writing a 400 word essay about Buck‟s conflict in the story “The Call of the Wild” Therefore, all these essays will show the way each student organized their ideas to develop the discourse topic Secondly, student essays in each class were collected into a folder

by the class monitor, and essay collections of two classes were sent to their teacher

by email for scoring Then the teacher sent the results back to the students through email too It brings two advantages for the researcher One is the convenience for collecting all the data without difficulty and the other is that the essays were scored

by the teacher; therefore it is an appropriate set of data for searching findings for the research questions

2.2 Research methods

2.2.1 Topical Structure Analysis method

The choice of TSA as the research focus is motivated by the following considerations First, TSA probes an important aspect of texts, namely the patterns

of maintenance and shifts of topics Such patterns contribute considerably to the coherence of texts, to the identification of what a particular stretch of discourse is about, and, consequently, to the comprehensibility of texts Therefore, this is an objective method for analyzing the student writing Second, TSA allows for the quantification of data which makes the study more reliable In summary, using TSA

as a tool for analyzing student essays is carefully considered for the research purposes

2.2.2 Data collection techniques

The data are essays collected from the student written products of two classes

at ULIS, VNU They include the essay final drafts, and all of these written products were rated by their lecturer of Literature in English The rater is lecture Lê Thành Trung who has great knowledge on English in general and on American Literature

in particular

In the population of forty essays, twenty-six essays consisting of thirteen rated and thirteen low-rated ones are taken as a primary source for data analysis In the following part, I would explain how the samples of the research are selected

high-For the research objectives, it is

necessary to take the equal numbers of

high-rated and low-rated essays By having

a glance to the essay score description

(Table 1), it could be seen that half the

essays in population are the essays with the

scores of A and A-, the other are the essays

Table 1: Essays’ score description

While the total of A+ and A essays is 13 essays and the number of low-rated essays (B, B-, C+, D-, F) is 14 essays Therefore, to have a good balance of low-rated and high-rate essays for analysis, one essay mark B is taken out Therefore, two groups are selected, one group contains thirteen high-rated essays including two A+ and eleven A essays, and the other group is made up by thirteen low-rated essays including six F, three D-, one C+, two B-, and one B essays

In summary, the numbers of essays which are chosen for analyzing the difference between low-rated and high-rated essays in this research are thirteen good quality and thirteen low quality essays

2.2.3 Data analysis procedure

This research follows the coding guidelines for TSA proposed by Schneider and Connor (1990) (See in the Appendix 1, page II) to investigate the essays To analyze the essays, four steps are implemented The first step is to divide the essays into t-units The second step is to identify and mark all the topical subjects in the essays, and the third step is to define the topical structures of the topics in the clauses The last step is to construct a diagram corresponding to the topical structure

of the essay to identify the topical progressions of the clauses in the essays

The following examples will illustrate the way 26 essays are analyzed The examples are two first paragraphs of the essay written by a Vietnamese student

Identifying t-units and topical subjects in the essays

In general, the length of the essays will be measured in terms of t-units In other words, t-unit can be defined as the shortest unit which a sentence can be reduced to, and consisting of one independent clause together with whatever dependent clauses are attached to it Compound and compound-complex sentences

which consists of two clauses connecting by the conjunctions and, or, but, for, nor, yet would be viewed as containing two t-units Predicates that share the same

subject are taken as one unit with the subject omitted The first step in conducting TSA is to number the clauses in the paragraph As in the examples, the first paragraph consists of three clauses, and they were numbered from 1 to 3

Paragraph 1:

(1) The call of the wild (T1) was the best seller book which brought fame to Jack London in

1903 (2) The book (T1) tells a trip of a dog in his life going through a lot of happiness, hardships and challenges (3) When Buck becomes a sled-dog with a struggle for survival, he (T4) must be used to the law of club and fang day by day, lesson after lesson

And the second paragraph has 10 clauses, and they continue to be marked from 4 to 13 at the beginning of the clause

Paragraph 2:

(4) In order to live in a wild life, Buck (T2) has an external conflict with Spitz- another sled-dog during their trip to go to Northland (5) The first reason (T1) leads to conflict is when Buck saw the violent fight between Curly and a husky dog, (6) it is the moment (T3) that Buck aware of the law of wild life: you must be a winner or you must die (7) In addition, it is also the moment (T5) that Buck recognizes the wildness, cold, and ruthless in Spitz‟s face- “ Spitz run out his scarlet tongue in a way he had of laughing” (8) and Buck (T2) hate him with “bitter and deathless hatred” (Source) (9) The second cause (T1) is that Spitz is very domineering; (10) he (T1) always “growled sharp reproof now and again or cunningly threw his weight in the traces to jerk Buck into the way he should go” (11) The next and the most crucial reason (T1) leads to the long fight between Buck and Spitz is that his nest was occupied by Spitz after he had a meal (12) and “till now, Buck (T2) had avoided trouble with his enemy, (13) but this (T2) was too much”

Identify the topical structures of the clauses in the essays

The second step is to identify the topical structures or the combinations of ISE, mood S, and topical S of the t-units in the paragraphs The ISE, topical S and mood subject of each t-unit in the paragraphs are identified and then the topical structure

of each clause are defined by adding T1, T2, T3, T4 or T5 after the bolded topical S

of that t-unit Following are the examples of two paragraphs above after they are analyzed

Paragraph 1:

(1) The call of the wild (T1) was the best seller book which brought fame to Jack London in 1903

(2) The book (T1) tells a trip of a dog in his life going through a lot of happiness, hardships and challenges (3) When Buck becomes a sled-dog with a struggle for survival, he (T4) must be used

to the law of club and fang day by day, lesson after lesson

Paragraph 2:

(4) In order to live in a wild life, Buck (T2) has an external conflict with Spitz- another dog during their trip to go to Northland (5) The first reason (T1) leads to conflict is when Buck saw the violent fight between Curly and a husky dog, (6) it is the moment (T3) that Buck aware of the law of wild life: you must be a winner or you must die (7) In addition, it is also the moment

sled-(T5) that Buck recognizes the wildness, cold, and ruthless in Spitz‟s face- “ Spitz run out his scarlet

tongue in a way he had of laughing” (8) and Buck (T2) hate him with “bitter and deathless hatred” (Source) (9) The second cause (T1) is that Spitz is very domineering; (10) he (T1) always

“growled sharp reproof now and again or cunningly threw his weight in the traces to jerk Buck into

the way he should go” (11) The next and the most crucial reason (T1) leads to the long fight

between Buck and Spitz is that his nest was occupied by Spitz after he had a meal (12) and “till

now, Buck (T2) had avoided trouble with his enemy, (13) but this (T2) was too much”

After all the topical structures of the clauses are identified, the frequency and proportions of T1, T2, T3, T4, and T5 are calculated for further analysis

Identifying the topical depth of the subtopics in the essay

Topical depth is the hierarchy of the subtopics in the student essays In another words, the topical depth shows the order of the topical subjects appearing from the beginning to the end of the essay

In order to identify the topical depths of the t-units in the paragraphs, all the topical Ss of the t-units are charted into a diagram In the diagram, the topical depth

of the topical subjects in the essays are numbered from 1 (the highest subtopic - the first topical subject appearing in the essay) and increase if there is one more subtopic appear (the lower ones) When the next subtopic is the same as the previous one, that subtopic comes back to the topical depth of the previous subtopic

The arrangement of these subtopics in the diagram will be discussed clearly in this paragraph Having a look at paragraph 1, the first and the second t-units have the same topical subject (the call of the wild and the book-having the same referent), these subtopics are in the same topical depth (number 1), and the progression used in these t-units is parallel In the t-units (2) and (3), the topical subject of the second t-unit (the book) is different from the topical subject of the

second t-unit (Buck), the second t-unit is in the higher topical depth than the third one, and the third subtopic is in the topical depth (number 2) In addition, the change of topical depth is a sign to identify the topical progressions of t-units When two or more consecutive t-units are in the same topical depth (1) and (2), (6) and (7), they belong to parallel progression When the topical subject moves from higher topical depth to the lower one as in t-units (4) and (5), (5) and (6), (8) and (9), (9) and (10), (11) and (12), the development of these t-units belong to sequential progression When the topical subjects of the t-unit moves from higher topical depth to lower one and then come back to the higher one as from t-unit (4) to (8), and from (8) to (12), the progression of these t-units is extended parallel progression

Identify the topical progressions of the topical subjects in the essay paragraphs

After the topical Ss are charted, the topical progressions of each paragraph are counted Sequential topics are indented progressively A topic with EPP is lined up under the parallel topic to which it refers When a chart is made to show the topical structure of a text, the progressive indenting represents topical depth

From the diagram, it can be seen that paragraph one above contains one PP (from clause 1to 2), and one SP (from clause 2 to 3) Paragraph two is more

complicated, it consists of one PP (from clause 6 to 7), five SPs (from clause 4 to 5,

5 to 6, 8 to 9, 9 to 10, and 12 to 13), and three EPPs (from clause 4 to 8, 8 to 12, and

9 to 11) After the topical progressions of all paragraphs are identified, the numbers

of PP, SP, and EPP are counted for further study

In this chapter, I have presented the research design and methodology As can

be seen the research approach is the quantitative analysis of the student essays The research applies TSA method for identifying the physical features and topic development of essay paragraphs written by Vietnamese students In the population

of forty essays of two classes E11QH2009 and E12QH2009 written in their Literature in English course in the year of 2013, thirteen high-rated and thirteen low-rated essays are selected to examine the physical features and the portions of the topical progressions of the introductory, body and concluding parts of two groups

CHAPTER 3 - FINDINGS AND DISCUSSIONS

3.1 The physical features of essays

3.1.1 Findings on the essay physical features

3.1.1.1 Numbers of essay t-units, paragraphs and sub-topics

Overall, the physical features of the essays revealed the difference between low quality and high quality writings

Table 3 below illustrates the mean (the average value of t-units per an essay), mode (the most frequently occurred value of t-units), and the median (the middle value of t-units) of the numbers of t-units, paragraphs and topical depths in the essays From the statistics in table 3, the physical difference between low-rated and high-rated essays did not show in the numbers of paragraphs per essay but in the numbers of t-units and sub-topics In terms of numbers of paragraphs, most of high-rated and low-rated essays contained four paragraphs, which were one introductory paragraph, two body paragraphs and one concluding paragraph In general, these essays had the same rhetoric pattern which was the general way discourse topic is developed

low-rated high-rated low-rated high-rated low-rated high-rated

quality essays consisted of about 29 t-units in an essay However, the mean, mode, and median of high-rated essays were 25, 23, and 24 respectively It showed that the numbers of clauses in high quality essays was just less than 25 which were fewer than low-rated essays In topical depth‟s analysis, the difference between the average number of topics in low-rated and high-rated was not large, about two topics, but they significantly differed in terms of range which was the distance between the highest and lowest numbers of topics per essay While the range of low-rated essays was 10, the high-rated essay range was just 5 It meant there was a large distance in topical depths of the low-rated essays (10 topics)

In summary, high-rated essays and low rated essays were different from each other in terms of t-units and topical depths

The numbers of t-units in the whole essays, and in different parts of the essays

In previous part, general finding on the difference between high and low rated essays was found; this part examines the difference between low-rated and high-rated essays in terms of the numbers of t-units in different parts of the essays Table 2 describes the mean, mode, and median of the clause numbers in introductory, body and concluding paragraphs of thirteen high-rated and thirteen low-rated essays

Introduction parts Body paragraphs Concluding paragraphs

High-rated low-rated High-rated low-rated High-rated low-rated

essays In addition, the mode and the median in high-rated essays are 3, which indicate that most of the essays consist of three t-units in the introductory paragraph In low-rated essays, the mode and the median are 3 and 4 respectively, which indicate that there are more than half of the low-rated essays consisting more than 4 t-units per introductory paragraph

In summary, the length of introductory paragraphs significantly contributes to classify the quality of two groups

3.1.1.2 Findings on the topical structures of clauses

Figure 1: The topical structures of clauses in low-rated and high-rated essays

The figure describes the topical structures of clauses in thirteen high-rated and thirteen low-rated essays written by Vietnamese students The statistics indicates that there is no significant difference between two groups in terms of topical structures In both groups, topical structure type 1 (T1), and topical structure type 2 (T2) are dominant in student essays Topical structure type 1, which the ISE, mood subject, and topical subject of most clauses coincide, is slightly greater in high quality essays than in low quality ones with the proportion of 43.1% and 40.3% respectively Topical structure type 2, which topical subject and mood subject coincide while ISE is separate is greater in low quality essays (43.5%) than in high quality ones (41.8%) Topical structure type 3, which ISE and mood subject

43,1

41,8

8,3 2,2

4,6

High-rated essays

T1 T2 T3 T4 T5

coincide while topical subject is separate, type 4, which ISE and topical structure coincide while mood subject is different, and type 5, which ISE, topical subject and mood subject are all separate, account for a small portion of the topical structures used

In summary, the topical structures did not reflect the difference between rated and high-rated essays

low-3.1.2 Discussion on the physical features of essays (Research question 1) 3.1.2.1 Numbers of t-units, paragraphs, and sub-topics

From the finding‟s part, it can be concluded that physical differences distinguish between low-rated and high-rated essays

In terms of paragraphs, the high-rated and high-rated essays are similar in the numbers of paragraphs per essay, which indicates that they have the same rhetorical pattern, which included four paragraphs, one introduction, two body paragraphs and one concluding paragraph The introductory paragraph introduces the discourse

topic which is the conflict in the novel “The call of the wild” The first body

paragraph describes the conflict in the novel, the second body paragraph show how the conflict is solved The last paragraph concludes what has been discussed in the body paragraphs In short, there is no difference in rhetorical pattern of the low quality and high quality essays

In terms of t-units, low-rated essays contain more t-units than high-rated essays At low-rated essays, t-units are often short, and the ideas are expressed by many clauses Although the requirement of the topic is to write an essay in 400 words, most of the low-rated ones exceed the required word number In this part, the discussion focuses on the difference between two groups of essay quality in the numbers of t-units in three main parts: introduction, body paragraphs and conclusion As far as an essay organization is concerned, both English and Vietnamese essays share these three components: introduction, body and

Vietnamese indirect ways of expressing ideas decrease the level of coherence of the essays, especially in introductory and concluding paragraphs

The low-rated essays also contain more sub-topics than high-rated ones The students who wrote high quality essays introduced fewer new subtopics in their essays than did the low proficiency group High-rated essays, writers have the ability to develop the topics in their compositions more evenly cross several topic levels than writers of low quality essays Good writer tend to return to higher topic level at the end of the composition rather than low rated essays

3.1.2.2 The topical structures of t-units in essays

Overall, the tendency in the uses of topical structures which indicates the realization of the ISE, mood S and the topical S of the clauses in high-rated essays

is similar in low-rated ones Two groups mainly use the topical structure type 1(T1) which the ISE, the mood S and the topical S coincide and topical type 2 (T2) which the ISE is separate while mood S and topical S coincide to develop the ideas in the clauses The large proportion in the use of topical structure type 1 suggests that students often represent their ideas in a direct way by starting the clause with the topical subjects Especially, the topical structure type 1 often uses in the first clause

of a paragraph in introduction and body paragraphs

3.2 Proportions of topical progressions in essays

This part aims at investigating the relationship between holistic evaluation and the topical progression used in essays written by Vietnamese students This part focuses on comparing and contrasting between low-rated and high-rated essays in terms of the topical progression occurrences in the essay paragraphs In particular, topical progression portions of introductions, body paragraphs, and conclusions of two groups are examined to have a deeper view on the contribution of each type of progression on the quality of the essays

3.2.1 Findings on the difference between two groups of essays

3.2.1.1 Topical progressions used in all essay paragraphs

Comparing and contrasting the progression portions of all paragraphs

In general, there were significant differences in the proportions of each type of progressions between high-scored and low-scored essays

Figure 2: The difference on the portions of each type of topical progression in Low-rated and high-rated essays

Figure 2 indicates the proportions of different progression types in two groups

of essays According to Figure 2, the proportions of topical progressions used in the highly-rated essays are different from those used in the low-rated ones; especially these two groups differ in the proportion of SP and PP

In general, SP are dominant in both high-rated and low-rated essays, following

by the PP and EPP respectively Portion of SP in high-rated essays is greater than in low-rated ones SP accounts for 58.3% in high-rated essays, and 43.9% in low-rated ones The proportions of EPP in two groups are quite similar with 13.9% in high-rated essays and 12.8% in low-rated ones The most significant difference in two groups is in the proportions of PP The low-rated essays contain more PP(s) than the high-rated essays While the portion of PP in low-scored essays is 38.3%, the portion in high-scored was 27.8%

27,8

13,9 58,3

High-rated essays

38,3

12,8 43,9

Low-rated essays

In summary, the difference in the proportions of SP and PP makes the difference in the essay quality The high-rated essays contain more SP and less PP(s) than low-rated essays

Comparing and contrasting the topical portions of body paragraphs

Obviously, different proportions of different type of topical progression found

in two groups of high-rated and low-rated essays However, it is important to note that each part of the essays including introductory, developing and concluding paragraphs has specific roles in the essays, and their roles partly affect the use of topical progressions in their paragraphs After describing the features of all essay paragraphs, this part continued to reveal the statistic on the differences between two groups in body paragraphs The finding on developing paragraphs of essays was shown in this part, because the topical progressions of body paragraphs contributed

to most of the progressions of the whole essays

Figure 3 indicates the

proportions of types of progressions

in the development paragraphs of

student essays Overall, differences

between low-rated and high-rated

essays in the body parts are the same

as the whole essays Figure 3: The proportions of different types of

progression in the body paragraphs of two groups of essays

Like the tendency of the progressions in all paragraphs in the essays, the features of body paragraphs include the high proportion of SP (57%) in high-rated essays and small proportions of PP (27%) and EPP (16%) In contrast, low quality essays contain a large proportion of PP (41.8%) while SP is 43.4%.In addition, there are balanced uses of PP and EPP in high-rated essays Balanced uses meant the gap between percentages of PP and EPP is small, as in the figure the ratio of PP and EPP is just 27/16, while it is 41.8/14.8 in low rated essays

41,8 14,8

43,4

27 16

57

Parallel Extended parallel

Sequential

high-rated essays Low-rated essays