An Empirical Study on the Application of Theme Theory in the Field of Writing Pedagogy

English writing instruction is an important part in college English pedagogy. Traditional way of teaching English writing lays much emphasis on word, grammar and sentence rather than the level of discourse. Under the traditional way, the students have difficulties to yield well-organized and coherent compositions. Theme Theory provides a theoretical foundation for a new method of teaching English writing. Thematic structure of the text is closely with the textual organization and the application of Theme Theory in teaching English writing helps improving students’ writing coherence and overall writing proficiency. This paper gives an experiment on the application of Theme Theory in teaching English writing and attempts to vertify the validity of the application of Theme Theory in the writing pedagogy of our college English.

A coherent text at discourse level largely depends on the semantic relations between sentences, such as thematic structure. In Fries' account, a good written text is one where the thematic content correlates with the structure of the text taken as a whole and with the method of development chosen. The role of theme in a good text rests on the close association between the grammatical subject and the topic of the sentence. This paper applies Theme Theory in Systemic-functional Grammar to deal with textual coherence and to illustrate that thematic progression is a major contributor to textual coherence and that the application of Theme Theory in English writing teaching will improve students' writing competence. The chief purpose of this paper is to examine how thematic structure affects the quality of writing on the one hand, and to testify the validity of the instruction on Theme Theory in English writing teaching on the other.

The Studies on the Concept of Theme and Rheme
Following the viewpoint of psychological foundation of syntactic construction, the terms theme and rheme are put forward to psychology communication. Mathesius (1939) describes word order in Czech and other Slavic languages. He defines theme as the starting point of utterance and rheme as the core of the utterance. And then he, in 1940, defines theme as the foundation of the utterance which is being said about in the sentence, and rheme as the core which refers to what the speaker says about the former.
Mathesius' very purpose to make researches on theme and rheme is to analyze how information is distributed in the sentence and then to find out the communicative effect of the distribution of known or given information and new information in a discourse. According to him, theme always carries known information, and theme is known or least obvious in the given situation and the starting point from which the speaker proceeds, while rheme is always the carrier of new information, so it is called the core.
Mathesius' research on theme and rheme exerts a great influence on later linguists.
Halliday, the founder of Systemic-functional Linguistics, makes significant contribution to the development of the thematic study. Halliday (1985), in his An Introduction to Functional Grammar, divides theme into three types--simple theme, multiple theme and clausal theme.
Simple theme: If the theme of a sentence consists of just only one structural element, it is called simple theme. E.g.: The duke (T) // has given my aunt that teapot (R). (Halliday, 1985: 38) Multiple theme: It refers to more than one structural element as the theme of a sentence. Multiple theme could include three structural elements, which are respectively called topical theme (experiential theme), interpersonal theme and textual theme. The three kinds of themes correspond to the three components of Halliday's theory of metafunctions: ideational, interpersonal and textual. Halliday (1985) holds that a sentence is, at one and the same time, a representation of experience, an interactive and a message. E.g.: Unfortunately (T-Int), however (T-Tex), the 'Un-artist' (T-Exp)// proliferated within the art instructions as well (R). (Thompson, 1996: 137) Clausal theme: A clause serves as theme in the complex sentence, which is constituted by a head (dominant) clause plus a modifying (dependent) clause, and this clause is called clausal theme. E.g.: If winter comes (T)// can spring be far behind (R)?
Halliday also divides Mathesius' concept of theme into two distinct systems: thematic system and information system. However, he holds different views about the relation between theme-rheme structure and given-new information structure from Matheius'. He states that:  listener-oriented. (1970: 299) In most cases, there is a parallel equivalence between theme and given on the one hand and between rheme and new on the other. However, there is no perfect correlation between them. Given and new information is context based. One can only decide which part of a message is new and which part is given within a linguistic or situational context. In English, themes representing new information usually appear in the following conditions: a. complement of a nominal group as theme in declarative clause; b. theme in cleft clause (predicated theme); www.ccsenet.org/elt English Language Teaching Vol. 6, No. 5;2013 c. theme in wh-clause.
Other descriptions on the concept of theme in general and its place in the theory of Systemic-functional Linguistics can be found in Jimenez Julia (1986), Downing (1991), Martin (1995), Eggins (1994), Kuno (1972), Downing and Locke (2002), among others. Downing (1991) also discusses marked and unmarked theme in different mood structures, and concludes that, although the thematic element in all of them is the point of departure, is not always "what the clause is about". Because theme is identified with first position in the sentence, and because English, in the unmarked case, has a strict word order. E.g: 1) Subject in a declarative clause The Gauls (T) sacked Rome. (Downing, 1991: 123) 2) Finite + S in a polar interrogative Did (T) the Gauls sack Rome? (ibid) 3) Wh-element in a Wh-interrogative What (T) did the Gauls do? (ibid)

4) Predicator or Let's in an imperative
In the case of marked word order, that is, when we have a fronted Complement or Adjunct in the clause, the fronted element will be the theme of the clause, as in the following examples.

1) Fronted Adjunct
In 390 Downing (1991) distinguishes between the theme of a clause and its topic, and points out that not all ideational themes are topical themes.
The discussion so far has concentrated on the many views and definitions of the notion of theme and rheme in general. Next, the author provides an account of Thematic Progressions.

Thematic Progressions
F. Danes put forward four models of Thematic Progression (TP) and claims that they can be used in various combinations in any discourse. By the term "Thematic Progression", he means "the choice and ordering of utterance themes, their mutual concatenation hierarchy, as well as their relation to the hyper-themes of the superior text unit (such as paragraph, chapter, etc.), to the whole text, and to the situation. Thematic Progression might be viewed as the skeleton of the plot". (1974: 114) Thematic progression is concerned with relatedness: where themes come from, and how they relate to previous themes and rhemes. Danes considers thematic progression to be one of the representations of connexity in a text: one of the ways in which a text displays coherence.
The other scholars have also confirmed the important role TP plays in constructing coherent texts. For example, Angela Downing (2001) points that the essential nature of a text lies largely in terms of its semantic coherence and the reason for less than optimal coherence may be that the writer is simply not good at controlling the mass of new information that is successively accumulated as the text unfolds. This mass of information is mostly so extensive that the writer should make a choice, and this choice, is determined directly or indirectly by the selection of utterance Theme.  Vol. 6, No. 5;2013 On the basis of Danes' models, Huang Guowen (1988) generalizes six models, all of which frequently appear in the ordinary use of English. Dubois (1987) performs an extensive analysis of a corpus of biomedical slide talks, and updates the typology to include a few new types, which reflect the fact that some progressions are gapped, that is, the Theme is picked up from an utterance that is not immediately preceding. He gives gapped variations of Danes' models. A gap takes place where there is intervening material between the theme and its link to the previous discourse.
Based on the above research on TP patterns, the author decides to concentrate on the following six TP patterns for the study: 1) Linear Progression: The rheme in the preceding clause becomes the theme of the subsequent clause.
2) Constant Progression: The theme in the preceding clause is employed as the themes in the subsequent clauses.
3) Focused Progression: The rheme in the preceding clause is continuously employed as the rhemes of the following clauses.
4) Derived-theme Progression: The first clause is the topic sentence; the themes in the subsequent clause may derive from the theme of the first clause.

5) Split-rheme Progression:
The themes in the subsequent clauses may derive from the rheme of the first clause.
6) Summarized Progression: The theme in the last clause is the same as the theme in the first clause. In this case, the main idea is often emphasized in the last clause.

Theme Theory and Coherence
According to Longman Dictionary (2000), coherence is natural or reasonable connection, esp. in thought or words. Halliday (1976) pointed out that words cannot be formed by jumping from one point to another point disorderly. Instead, the words must develop regularly and reasonably according to the coherence of a theme and the probability of the words development. Halliday (1976) defined coherence as the probability of connecting what has been mentioned above. To ensure the successful development of a topic, the application of various techniques for cohesion is vital. Some forms of cohesion are realized through the grammar and others through vocabulary (Halliday, 1976). Cohesion contributs to coherence. Thematic progression is a kind of cohesion (Halliday, 1976).
Many scholars have noticed the relationship between Theme Theory and coherence. Danes (1974) proposed the thematic progression and applied it to discourse coherence. He emphasized the important constructional role of the Theme in the organization. He claimed that the degree of coherence in the text is reflected in the extent of the continuity of the thematic progression, which is reflected by the connection of similar language units. The discontinuity of this connection will lead to the interruption of coherence.
In Discourse Analysis, Brown and Yule (1983: 133) interpreted Theme's function as: connecting back and linking into the previous discourse, to maintain a coherent point of view; serving as a point of departure for the further development of discourse. Baker (2000: 121) also ponited out that Theme has two functions: (a) it acts as a point of orientation by connecting back to previous stretches of discourse and thereby maintaining a coherent point of view and, (b) it acts as a point of departure by connecting forward and contributing to the development of later stretches Halliday (1985: 67) explained that thematic progression is of crucial importance to the internal relation between sentences in the text. Thematic progression contributes to text coherence in that it lets us know what the text is about and where the text is going. By analyzing the thematic structure of a text, "we can gain an insight into its texture and understand how the writer made clear to us the nature of his underlying concerns".
In Textual Cohesion and Coherence, Hu (1994: 145) argued that the repeated connection of Theme and Rheme is one of important ways to contribute to the discourse cohesion and coherence.
Zhang & Liu (2003: 7-33) carried a study on the relationship between thematic structure and coherence and pointed out: "(a) theme choice locally connects sentences semantically; (b) as a whole, it helps the text form a whole integrity in the way of information organization; and (c) it helps to link the text and context."

The Studies on the Application of Theme Theory in English Writing
In recent years, researchers and language teachers have endeavored to locate connections between the application of Theme Theory and the coherence of writing. Rose David (2001) points that a text will lose its www.ccsenet.org/elt Vol. 6, No. 5;2013 coherence if the thematic patterning is destroyed, and studies of thematic selection have important pedagogical application and it seems possible to teach students about theme quite explicitly, perhaps beginning with the second and third of Danes' patterns. Angela Downing (2001) proposes that students' compositions might be improved as regards to what is subjectively termed the "flow" of the text, by consciously selecting certain thematic progression patterns. Downing and Locke (2002) hold that thematic structure might take a prominent part in writing. Sarah North (2005) researches into the disciplinary variation in the use of Theme in undergraduate essays in Open University by analyzing students' written materials in history and science. Taking thematic choices as a point of discussion, he comes to the conclusion that different knowledge disciplinary backgrounds can make impact on students' application of different theme categories, which would further influence the structure and information organization of students' writing. Moorse (2008) from Etisalat University uses the linguistic functions of Theme-Rheme and Given-New to show how the writer can easily arrange the clauses to allow the reader better access to information and reference to clearly mark what is assumed to be known by the reader.
Some Chinese researchers have also confirmed the positive effects of Theme Theory in improving students writing. Liu Chendan (2001) claims that teachers should guide their students to make a proper arrangement of the thematic structure in their writing in accordance to genre, such as giving an entitled composition, offering an outline, asking students to write a summary, etc. Fang Li (2004) employs patterns of TP to analyze the English writing by Chinese learners with a purpose to investigate their patterns of text organizations from both micro level and macro level. Then she conducts a comparison between the text structures by native English learners and Chinese students. Xu Yuchen (2006) probes the theory into the study of advertisements by analyzing the thematic structure and information structure. From his point of view, theme focus and information focus are helpful for us to interpret the information placement of a text. Besides, he points out that the English teachers' awareness shall be called on to the information placement in English discourse when they teach reading and writing so as to help students read better and write better. Zhao Nan and Zhou Xiaomo (2008) randomly select ten argumentative writings of different length as corpus from English newspapers, journals or websites and study the features of constructions of argumentative writings. They find that most of the themes in argumentative writings are unmarked, which are of clear distinction from those in descriptive and narrative writings. Their study is helpful to the teaching of EFL writing.
So far, the guiding role of Theme Theory in the field of writing pedagogy has been recognized by some scholars. As far as the focus of this research is concerned, more investigations should be carried out in the direction of text organization of EFL writing at the discourse level due to the relative ignorance of this area at present. So based on the previous studies, the present paper gives an experimental study on the application of Theme Theory in the field of writing pedagogy and suggests how Theme Theory might be of practical use in EFL classroom.

Subjects
The experiment was conducted in the English Department of China Three Gorges University (the author's university) in 2011. Two classes of English majors of Grade Two were chosen. The total number of the subjects was 50. Each class had 25 students. Class One was taken as the experimental class and Class Two as the controlled class. They had similar learning background. They both had years of English learning and English writing practice. And there was no obvious distance among the scores of their entrance examination (to the university), as Table 1 shows, which was analyzed through the software of SPSS 13.0. So English level of these two classes was familiar and there was no significant difference between them.

Research Question
This experiment focused on the following question: www.ccsenet.org/elt Vol. 6, No. 5;2013 Does the application of Theme Theory in English writing pedagogy help improving students' writing coherence and their overall writing proficiency?

Procedure
The experiment was carried out in the normal classroom teaching in the 3 rd semester of English majors. The subjects had the same course book and followed the same syllabus. The experiment took up 20 class hours. That is, according to the arrangement of 2 hours of writing classes a week, the experiment lasted 10 weeks. To avoid variation in teachers, the same teacher (the author) was in charge of the instruction of writing for the two classes alternatively.
In the first week, a pre-test on English writing was given. The topic "A Letter to the University President about the Canteen Service on Campus" was taken from the writing section of CET-4, in January of 2005. CET-4 was initiated in 1987 and continues until now and is held twice every year. CET-4 is a standard test as the nation-wide College English Test in China. So it is one of the most authoriatative and extensive examinations in China. The two classes were required to finish it within 30 minutes in accordance with the requirement of CET-4 writing. Then the author graded the compositions from the perspectives in reference to the scoring criteria of CET-4, content, organization, coherence, grammar, language and norm of writing. The author used the full mark of 100 in the grading for the convenience of statistical analysis. After the grading, the number of themes and TP patterns in the compositions was calculated.
In the next five weeks, from week 2 to week 6, the author taught English writing to the two classes in different methods. The controlled class still received instruction in the normal way, which focused on vocabulary, grammar, syntactic structure, diction, etc The experimental class, besides receiving the normal instruction, was systematically guided to some skills of producing coherent writing, such as thematic structure and TP patterns. Considering Theme Theory is new and difficult for the university students, the author gave the students adequate time to learn it. The author spent two weeks on the instruction of theme and rheme and three weeks on thematic progression. In the process of teaching in the experimental class, firstly, the author guided the students to master preliminary knowledge of Theme Theory, including definition of theme and rheme, their identification in different mood and the relationship between thematic structure and information structure. Examples of English were given to strengthen their understanding. Then the author introduced TP patterns to the students, and also guided them to analyze the model passages in thematic progression, finding what kind of thematic progression models are used in the passages.
The next three weeks from week 7 to week 9, was the period for consolidation and practice for the experimental class. In this period, firstly, the students were directed to do reading comprehension by the application of thematic analysis. Then the students learned to appreciate and evaluate model writing from CET-4 from the perspective of thematic structure. Meanwhile, the controlled class was guided to deal with the same learning materials in the normal way, such as discussing the topic sentence, the grammatical structure, new words, diction, etc.
In week 10, the last week, the author designed a post-test for the two classes. It's from the writing section of CET-4 in June of 1999. The topic is "Reading Selectively or Extensively". The test was finished within 30 minutes. Then the author graded the compositions with the same scoring criteria for the pre-test. The post-test compositions of the two classes were mixed up so that the author could not show bias in marking the students. After the grading, the number of themes and TP patterns in the compositions was also calculated.

Data Collection and Discussion
The author analyzed both pre-test and post-test through software of SPSS 13.0. For the purpose of the present study, only two items, coherence and total scores, were listed into the range of the statistical analysis. Table 2 is the writing performance of the two classes in the pre-test.  Vol. 6, No. 5;2013 123 The results in Table 2 show that there is no obvious difference in writing competence between the two classes.

A Survey on Pre-test
The following two tables are the distributions of themes and TP patterns in the compositions of the two classes.
In Table 3, we categorize theme into 3 types--simple theme, multiple theme and clausal theme, as mentioned early. Meanwhile, according to the function of elements in clause, we identify experiential (topical) theme, interpersonal theme and textual theme. When a clause only has one structural element as theme, the theme must be experiential theme (Halliday, 1985: 52). So we put simple theme and experiential theme together.  Table 3 and Table 4 reveal that there is no distinction in the distribution of themes and TP patterns between the two classes.

A Survey on Post-test
T-test results of the two classes in post-test are analyzed by SPSS, 13.0 and the data are presented in Table 5. From Table 5, we can see that the two classes have distinct difference on coherence and total scores. The experimental class has progressed greatly on both coherence and overall writing proficiency while the controlled class has only progressed a little. This proves that the students in the experimental class can achieve better writing after getting the instruction on Theme Theory.
To further confirm the function of the application of Theme Theory in improving writing, the author calculates the number of themes and TP patterns and investigates the distribution of themes and TP patterns in the compositions of the two classes. Both of the two classes use the three types of theme, but the distribution of the types of theme is different. The controlled class uses more simple themes than in the experimental class. This is not against our expectation, since simple theme is obligatory theme that contributes to the development of text. However, the experimental class uses more multiple themes and clausal themes. This shows that the experimental class is better at using different types of theme to achieve a variety of sentence structure. Another significant difference lies in the fact that the experimental class uses more textual themes than the controlled class. This fact reveals that the students of the experimental class are more skillful in making their compositions logical and coherent because textual elements do important cohesive work in relating the clause to its context. The linear and constant thematic progression patterns occur frequently in all compositions. This is not surprising, because Danes refers to them as basic patterns, and Dubois (1986) similarly characterizes them as canonical types, from which other patterns are derived. But obviously, the experimental class uses more TP patterns than the controlled class. Besides, the percentages of most patterns are higher in the compositions of the experimental class than those in the controlled class. The controlled class uses constant pattern most, which occupies 45.4% of the total percentage. The excessive predominance of this kind of TP pattern easily makes the texts dull. This shows that the students of the controlled class are deficient in organizing their texts effectively.
From the above analysis, we can see that there is a relationship between writing achievement and thematic structure of texts.

Samples' Analysis
In this section, the author presents two sample compositions with the same title "Reading Extensively" to explain www.ccsenet.org/elt English Language Teaching Vol. 6, No. 5;2013 how themes and thematic patterns work together through a text to signal its underlying unity and coherence. They clearly exemplify many of the features of the texts encoded by other students of similar language proficiency. For the convenience of analyzing, one paragraph is extracted from each sample composition.

Sample 1: A Good Composition
Reading (T1)  So long as we read the themes of this text, we know that the content of it concerns "reading", which is the starting point of the first T-unit. Then we can see the word "reading" is picked up again in different forms in the consequent text. We find that in the text "reading" is thematized 4 times, which amounts to 40% of the total themes. This fact reveals the author's purpose: he wants to draw the reader's attention to the topic of the text. The author also chooses the themes and varies theme types to avoid blankness caused by using the same type of theme. For example, the clausal themes in sentence 5 and sentence 6, the multiple themes in sentence 4 and sentence 10, the textual themes in sentence 3 and sentence 10, the interpersonal theme in sentence 10.
The TP patterns of the text are developed as follows: This text consists of 6 TP patterns: linear pattern (2 times), constant pattern (4 times), focused pattern (1 time), derived pattern (2 times), and summarized pattern (1 time). Theme 1 introduces the topic of the whole text. The TP patterns in sentences 2, 4, 7 and 10 can be assigned to be constant pattern with the same referent "reading" as their themes. In sentences 5 and 7 linear pattern is used. The rheme in sentence 10 is the same as that in sentence 1, and it is focused pattern. The theme is a part of the theme of sentence 2, so it is derived pattern. Sentence 9 also uses derived pattern, since the theme "Our minds" is derived from "The nutrition for our minds", the theme of sentence 8. The theme in the last sentence is the same as the theme in the first sentence. The main idea is emphasized in the last sentence. Summarized pattern is adopted.
As can be seen, the sentences of the text are linked together owing to the relationship of themes and rhemes of the sentences in the TP patterns. TP patterns make the whole text go smoothly and coherently. There are only one clausal theme (T5) and 2 multiple themes (T1, T3). Thus, the structure of the text tends to be simple. The word "reading" and its referent "it "are thematized 2 times. The rest themes are mainly personal pronouns except T5 and T6. Among the personal pronouns, "we" appears two times, "he" and "his", each of them, one time. These experiential themes lack logical connections internally. So it is difficult for the readers to grasp the topic of the text and the train of thought of the writer from the themes of the text.
The following is the TP patterns of the text: TP between sentences 1 to 2, sentences 3 to 4, sentences 6 to 7 is constant and sentences 7 to 8 is simple linear. The use of the TP patterns is much simpler than that of Sample 1. However, the main problem of this text is that some T-units and R-units are not involved in any progression. Especially, the T-unit of sentence 5 suddenly employs "Good students do a lot of reading" as the theme. So some sentences are not linked.
The above analysis has revealed that good writers are more skillful in organizing their compositions through the reasonable choices of themes and TP patterns.

Findings
From the results of the experiment, we find: 1) Writing coherence largely relies on the success of thematic structure of texts.
2) The significant progress on writing coherence contributes to the significant improvement on overall writing proficiency.
So the experiment gives a positive answer to our research question, that is, the application of Theme Theory in English pedagogy helps improving students' writing coherence and overall writing proficiency.

Summary of the Study
In this paper, we conduct an experiment to testify the validity of the application of Theme Theory in the writing pedagogy of our college English. By the experiment, we draw the following conclusion: The application of thematic structure has a positive effect on the students' writing competence and the instruction based on Theme Theory proves practical use in writing pedagogy.
The present study may help us to reflect on the inadequacies of our traditional way of teaching English writing and probe into a new way of writing instruction under the guidance of Theme Theory. This new method can help our EFL learners get out of the narrow scope of syntax and view the whole composition from the organization of text. However, this method is not widely recognized. So we do hope that the tentative study done in this paper could offer any possible insight into writing teaching as well as into the perfection of writing quality.

Limitation of the Study
Although the teaching method has been proved to be practical and effective by the results of both pre-test and post-test, there are some limitation of the study: Firstly, the teaching process in experimental class last only 10 weeks, which is a short period of time. The writing classes are so limited. The participants in this study limit to one major in one university. So, the www.ccsenet.org/elt Vol. 6, No. 5;2013 generalization of the findings has to be made with caution. Therefore, research with a longer time, a larger sample size and various subjects from diverse majors in different universities of varying places is to be expected.
Secondly, as the theory is complex for the undergraduates, they feel that it is not easy to master the theory and apply it to writing. As a result, it has been found that students' application of thematic progression patterns is a bit stiff and unnatural.. Some students tend to use one type of progression through the whole text. Sometimes they just try to apply patterns for the application itself and ignore the normal semantic coherence. As to this problem, future researches may design certain teaching plans that can help students to avoid such difficulties.
Since this research is still preliminary, limitations are inevitable. In spite of its limitations, this study design offers a practical method to assist writing teaching to some extent. It is necessary to carry out further research in this field.