Malaysian ESL Students’ Perception of Metadiscoure in Essay Writing 

This study investigated the use of metadiscourse markers and highlighted the problems that Malaysian ESL students face in writing essays. An action research approach using both quantitative and qualitative analyses were used to explore metadiscoure usage. The subjects of this study were diploma students who had enrolled in an English for Academic Purposes (EAP) course in a private college in Malaysia. A total of 143 scripts of the essays written in the essay writing section of the final examination were assessed and analyzed using Hyland’s (2005a) interpersonal model of metadiscourse. Results revealed that transitions were the most frequently used interactive metadiscourse, making up 54% of the total interactive markers, while evidentials were the least frequently occurring with only 2 occurrences. Under the interactional resources, engagement markers were found to be the most favored, accounting for 40.7% of the interactional markers, while attitude markers appeared the least with only 3 occurrences. The findings of this study suggest that metadiscourse practices play a crucial role in creating a cohesive and a coherent piece of writing. In addition, ESL students need to be explicitly instructed on the communicative function of metadiscourse markers in academic writing.


INTRODUCTION
Discourse and text analysis have been the areas of interest among linguists for many years, because of their contributions to the identification of how language is used in various academic genres such as in research articles and academic lectures. One way to study the language selection of academic and disciplinary writers is through metadiscourse knowledge. Hyland (2000) defines metadiscourse as "the linguistic resources used to organize a discourse or the writer's stance towards either its content or the reader". Other scholars have also provided their definitions of metadiscourse. To Williams (1981), metadiscourse is "writing about writing, whatever does not refer to the subject matter being addressed" (p.212). For many others, they are linguistic features which signal the presence of the writer in the text and guide the readers through the text by organizing, interpreting and evaluating the information (Vande Kopple, 1985;Mauranen, 1993). Metadiscourse focuses only on establishing a connection between a writer, readers and a text, and do not seem to add any information to the propositional contents.
In general, metadiscourse analysis shows how writers project or situate themselves into their argumentation in order to control their interaction with readers, and to signal their evaluation and commitments to a proposition through various lexicogrammatical features (Hyland, 2005b). Central to their importance, metadiscourse markers can enable writers to transform a mundane and difficult text into something coherent and reader-friendly. In addition, the writer's personality, style, voice, credibility, sensitivity to audience and relationship with the text also comes through in the use of metadiscourse markers (Hyland, 2000). Hyland (2005a) proposed the interpersonal model of metadiscourse based on the revisions of the earlier models (Hyland, 1998;2000). He clearly distinguishes between interactive and interactional metadiscourse. The interactive dimension serves to shape and limit a text to meet the needs and expectations of potential readers, as well as to navigate them through the text. The use of interactive metadiscourse enables writers to organize their discourse and arguments in a way to help readers understand the propositions through five subfunctions of transitions, frame markers, endophoric markers, evidentials and code glosses. On the other hand, interactional metadiscourse seeks to engage and pull readers along with the text and interact with them by imposing and commenting on their own message. That is, through interactional metadiscourse, authors express their textual voice in a way to open a dialogue and create a space for potential disputes from readers. Interactional metadiscourse has five resources which are hedges, boosters, attitude markers, engagement markers and self-mentions. All these markers enable writers to evaluate a proposition, engage readers, express solidarity, and anticipate potential objections on the part of readers (Hyland, 2013).
The investigation of metadiscourse in essay writing is of great importance, as it is a means of achieving cohesion and coherence in writing. The importance of essay writing is further amplified as it is considered as one of the significant instructional methods in teaching English for academic and specific purposes in academic and professional settings. As Hyland (2004) notes, it is important for students to possess good writing skills in a variety of genres as it can help them establish career opportunities and create positive identities in making life choices. Therefore, in most language classrooms, students are taught how to write short research projects, journal reviews, case studies, and other academic and semi-academic text types. For most of these genres, students are required to formulate the information and structure their discourse in the form of an essay format. Albeit its importance, it is important to note that writing is not an easy skill to master. In fact, many English as a Second Language (ESL) students view writing as a challenging task. This is because, the process of writing involves the writer to chart language onto his or her own thoughts while at the same time taking into account the expectations of the reader. The issues of difficulty in writing is heightened at tertiary level, as students at this level are required to produce written texts that are matured, more complex and sophisticated enough to be equivalent to their perceived intellect (Ismail, Hussin & Darus, 2012). According to Hyland (2007), to achieve academic success, tertiary literacy requires students to be skillful in producing the write ups that involve sustaining arguments and synthesizing ideas. It has been reported that students who are non-native speakers of English commonly face difficulties in grammar and pragmatic meaning when it comes to writing (Rose, 1985). To Paltridge (1996), their limited English language proficiency makes it challenging for them to convey their ideas in a cohesive and coherent way.
Among the common linguistic difficulties in writing, cohesion seems to be the most prominent challenge faced by ESL students. The concept of cohesion is "basically a semantic oneit refers to relations of meaning that exits within the text and that define it as a text" (Halliday & Hasan, 1976, p. 4). It aids in comprehending how a written or spoken discourse operates or "hangs together" via "cohesive ties or chains that link the presupposing and the presupposed across sentence boundaries" (Halliday & Hasan, 1976, p. 8). Cohesion can be attained in writing when the writer uses suitable vocabulary and various grammatical classes such as pronouns, conjunctions, references, lexical substitutions, and academic lexical bundles). The use of these cohesive ties can signal a writer's linguistic competence. Linguistic competence here refers to the writer's competence in using appropriate grammatical and semantic rules, in which Hymes (1972) posits it as communicative competence. Such language use is deemed necessary to show the credibility of a writer and prove that one is a proficient language user.
Numerous studies have been done on essay writing and the challenges it presents, particularly to the ESL students. Alarcon and Morales (2011) conducted a study among first year Political Science and Computer Science students enrolled in expository writing and communication skills class to investigate the presence of grammatical cohesion in their argumentative essays. They found that cohesive devices indeed facilitate their writing process and establish strong connections between different parts of the text and sentences. They concluded that cohesive devices such as grammatical and lexical structures contribute to the meaning-making process of the text and play a key role in achieving cohesion in writing. The study carried out by Liu and Braine (2005) on Chinese undergraduate students revealed similar results. They found that there is a positive relation between the frequency of cohesive devices and the quality of students' essay. That is, the higher the occurrence of cohesive devices are, the better the quality of the essay is. It was also discovered that lexical phrases are the most frequently used cohesive device to make the sentences more meaningful. Besides, Sinicrope (2007) carried out a study on 4 undergraduate students enrolled in an advanced expository writing course, to highlight the ways that proficient L2 writers use cohesive devices in their essay writing. The analysis of data indicated that students who scored high marks in their essays used all the three cohesive devices, namely reference, conjunctions and lexical cohesion. It was concluded that cohesive devices enhance the quality of students' expository writing.
Ting and Tee (2008) conducted a case study among Malaysian Teaching of English as Second Language (TESL) students to examine their academic writing skills. Their results suggested that the lack of knowledge of cohesive devices among these students caused their ideas to sound underdeveloped within stages of their writing. Although students were able to construct various sentences, there was a serious lack of cohesion caused by inadequate use of modal verbs, conditional clauses, and connectors to establish clear links between different parts of their texts. Dueraman (2007) also conducted a comparative study between Malaysian and Thai medical undergraduate students on the use of cohesive devices in both narrative and argumentative essays. Both groups of students were asked to write 14 narratives and 14 argumentative essays in English based on selected topics. Findings indicated that both Malaysian and Thai students used more syntactic ties such as reference and conjunction than semantic ties such as reiteration and collocation in writing the two genres. However, reiteration was found to be more favored by the Malaysian students.
The review of the existing literature clearly supports the need to continue conducting such investigations. Although extensive research has been conducted on the role of cohesive devices in essay writing in the L2 context, little has been done to highlight the use of metadiscourse markers and their role in achieving cohesion in this genre, especially in an ESL context like Malaysia. Existing studies on metadiscourse markers focused extensively on other written genres like academic research articles. Abdi (2002) investigated the use of interpersonal metadiscourse markers in 55 conclusion sections of research articles in social sciences and natural sciences; and Khedri, Ebrahimi and Chan (2013) conducted a crossdisciplinary study on the use of only interactional metadiscourse Markers in the results and discussion sections of research articles across four disciplines (English Language Teaching, Civil Engineering, Biology and Economics). Therefore, the present study aims to answer the following research questions: 1. What are the most and the least common metadiscourse markers used in Malaysian L2 students' essay writing? 2. How do Malaysian L2 students perceive the communicative functions of metadiscourse markers in their essay writing?

Participants
The participants for this study were 143 ESL students from a private college in Malaysia who had enrolled in an EAP course. In order to ensure that these participants were of same English language proficiency, they were asked to complete an English Placement Test (EPT) prior to attending the course. Their age ranged between 18 to 25 years. They were in their first or second semester, and from different Diploma majors including Diploma in Retail Management, Diploma in Event Management, Diploma in Culinary Arts, Diploma in Pastry and Diploma in Hospitality Management. It is important to note that foreign students were excluded from this study because some of them were English as a Foreign Language (EFL) learners and this might have affected the results of the study. Therefore, only the Malaysian students who use English as a second language (L2) were selected for this study.

The Corpus
The corpus used in this study was 143 essays written in the final examination of an EAP course which lasted for 1 academic semester. The examination paper constituted 3 sections; reading comprehension, grammar and essay writing. Data for this study was collected from the essay writing section which required the students to choose a topic out of the 2 exploratory topics and compose an essay of at least 250 words. The corpus amounted to 41,327. Table 1 presents a description of the corpus used in this study. The Analytical Framework

Hyland's interpersonal model of metadiscourse
This study follows Hyland's (2005a) interpersonal model to analyze the use of metadiscourse resources in students' essay writing. Under the interactive resources, there are five metadiscourse markers, namely transition, frame marker, endophoric marker, evidentials and code glosses. Transition markers aim to express semantic relations between independent clauses. By doing so, readers are guided to make pragmatic connections between sentences of the passage. Transitions are characterized by the portrayal of additive, contrastive or consequential meaning between ideas. Examples of transitions are and, in addition, equally, however, and therefore. Frame marker as its name suggests, signals the transition of moves and steps in the text. It focuses on the schematic structure of the text by labeling, sequencing, announcing, and shifting in topic. Generally, authors rely on frame markers such as first, lastly, and next to show sequence; in sum, as an introduction, and to conclude to label; my focus, I seek to, and I argue here to declare the purpose of discourse; and back to, in regard to, and well to indicate a shift in topic. The third kind of interactive metadiscourse is endophoric markers which direct readers to other parts of the text where supplementary materials are available. In other words, endophoric markers are references to additional content and allow readers to recover the preferred meaning of the author. Examples of endophoric markers include as noted above, see Figure X, and refer to the next section. Evidentials help distinguish the outsourced content from the original text through expressions like according to X, Y states that, and X asserts that. Essentially, evidentials play a key role in ensuring that the text reaches out to readers with greater accuracy and in varied dimensions. As for code glosses, authors would carry assumptions about the readers' cognition by deciding on the degree of extra information to give them. In this manner, phrases such as in other words, defined as, such as, and namely are often employed as code glosses.
The second category of metadiscourse, interactional, also constitutes five subtypes which are hedges, boosters, attitude markers, engagement markers and self-mentions. As defined by Hyland (2005a, b), hedges are uncertainty markers used by writers to withhold commitment and to lessen their responsibility in expressing a proposition. Examples of hedges include modal verbs such as may, might, could as well as adverbs such as perhaps or possibly. Boosters, on the other hand, are markers used to emphasize certainty with respect to the writer's claims and assertions. Examples of boosters are in fact, definitely, obviously, and it is clear that. Attitude marker is a device used to express a writer's attitude or viewpoint toward the propositional content. This can be the writer's personal feelings of surprise, obligation, and disagreement (Hyland, 2005a, b). Attitude markers can be of different grammatical classes, such as attitude verbs (agree, disagree), adverbs (hopefully, interestingly), and adjectives (appropriate, remarkable). Writers use engagement markers to establish an explicit connection with readers and involve them in their arguments by using expressions like consider that, note that, you can see that. Lastly, selfmention is represented by the use of personal pronouns and their possessive form such as I, we, my, and our, to make an explicit reference to writers in order to demonstrate their personal representation in their research (Hyland, 2001;2005a).

Procedure
This study adopted both quantitative and qualitative approaches to identify and analyze the frequency and functions of Metadiscourse markers in Malaysian ESL students' essay writing. A number of steps were taken to carry out this study. Firstly, the data were collected and manually analyzed using a top-down approach. This mean that all the transcripts were manually analyzed to retrieve all the lexicogrammatical features which function as metadiscourse. In the qualitative phase, all the occurrences of each marker were investigated in their context to determine and classify its main discourse function. To ensure inter-rater reliability, another rater checked all the occurrences of each marker and any misalignment was corrected upon reaching a consensus after checking each context carefully. Lastly, the examples were extracted from the original scripts to discuss the use of each of these metadiscourse markers and see how the ESL students perceived them in writing their essays.

RESULTS AND DISCUSSION
As Table 2 indicates, the total number of metadiscourse markers used in students' essays were 1147, with interactive metadiscourse markers appearing to be more favored (60.6%) than those of interactional (39.4%). The higher frequency of interactive metadiscourse shows that ESL students appreciated the importance of framing their discourse and guiding their readers throughout their essays, especially through the use of transitions and endophoric markers. By creating such metadiscursive connections, they seemed to be aware of the existence of potential readers. The use of interactional devices also indicates their evaluation, assessment, and degree of commitment to the given propositions while projecting themselves in discourse and engaging their readers.

Interactive Resources
As can be seen in Table 2, transitions were the most commonly used interactive metadiscourse in students' essays with 367 occurrences, which accounted for more than half of interactive markers used. As Hyland (2005a) notes, transitions are used to express semantic relations between clauses, and to establish logical connection between propositions. They can be portrayed in the form of additive (in addition), contrastive (however) or consequential meaning (as a result of), establishing a connection between ideas in the text. In the corpus of essays in this study, transitions were used to create a clear link between sentences, enabling a smooth flow to be formed throughout the essay in order to facilitate reader's understanding of the content. Some common transition markers used include and, or, in addition, furthermore, hence, thus, therefore, but, and also. However, it was found that transitions mostly functioned as additives in essays, comprising almost 65% of all cases.
1) Moreover, to enhance self-image, first impression always counts.
2) This is because every work place has their own uniform or dress code. It was evident that endophoric markers make up 29.3% of the total number of interactive markers, making it the second most frequent. Endophoric markers are typically used to refer readers to other parts within the text where additional information may be available. Sometimes, writers use endophoric markers to help readers grasp their intended meaning (Hyland, 2005a). In this study, it was found that ESL students mostly rely on personal pronouns such as it, they, them, and their, and demonstrative pronouns such as this and these to serve this function. In addition, giving reference in essays was all in the form anaphoric, in that, students only referred back to the previously mentioned word or idea in the text for its meaning.
3) If you do not want to turn off your phone completely, at least let it vibrate. 4) Student dress code is important in college as it is a great starting point for them to understand … and enhance their self-image. The third most frequently used interactive metadiscourse was frame markers with 93 occurrences, accounting for 13.3 % of all interactive devices. The expository nature of essays required the students to use frame markers in order to indicate the sequence in presenting ideas (5), to establish a setting or time frame relating to the topic of essay (6), and to frame the thesis statement and introduce the overall structure of the essay (7). It was depicted that frame markers mostly occurred in the first sentence of the introduction paragraph, functioning as a point of departure to help writers organize their discourse. 5) First of all, having an appropriate dress code in college will lessen students' wardrobe stress. 6) In today's society, students are often tied by abundance of rules. 7) In this essay, I will like to introduce the basic hand phone etiquettes to you. As for code glosses, analysis showed that they accounted for only 4.17% of total interactive markers. Code glosses were generally used by the students to elaborate on the topic either by providing examples or giving additional information. Giving further explanations and examples are considered interactive strategies to guide potential readers through the text and facilitate their understanding of the content.
8) Examples of sensitive topic are politic issue, culture of other races, and gossip about other people and so on. 9) First, the hand phone etiquettes in public is use hand phone in suitable place. For instance, hospital, libraries, meeting, police office and work. Finally, evidentials were found to be the least frequently used interactive marker, amounting to only 0.2%. The less occurrence of evidentials in the corpus may be attributed to the fact that students were required to write the essay in an examination setting, where they were unable to refer to external sources for information to include in the essay. In the following example, the student uses the evidential 'according to' to refer to the source of information that he has studied before as a way to support his idea: 10) Last but not least, according to the America Education Department, implementation of proper dress code in college will lead to decline of campus crime.

Interactional Resources
Results plotted in Table 2 indicate that engagement markers constituted the highest number of interactional metadiscourse, accounting for 40.7% of the total occurrences. Engagement markers are used to enable writers to form a relationship with readers by engaging them in the text. It is typically accomplished through the use of the second person pronoun 'you' and its derived forms 'your' and 'yourself', as well as a number of engaging verbs such as 'note', 'notice', 'consider', 'look', 'see'. However, it was found that Malaysian L2 students tended to overuse the pronoun 'you' (11), and were unaware of other grammatical forms to engage their readers in discourse. 11) When you go to a party, you do not wear the same clothes you have wearing before. Likewise, you dress for work, you will dress yourself to look like a professional. Self-mentions were found to be the second most frequent interactional metadiscourse. Selfmention is a conscious choice by writers to give explicit reference to themselves and show their stance (Hyland, 2005a, b). In the corpus of essays, it was depicted that students had a high propensity to use the inclusive noun 'we' and its possessive form 'our' in presenting their arguments. One reason for this may be due to the expository nature of the essay topics _college dress code and hand phone etiquettes_ which are related to students' lifestyle as they are college students themselves. Hence, they may have consciously or unconsciously inserted themselves and their experiences in discussing the topics, as in: 12) As a conclusion, hand phone etiquette is something that we must apply in our daily life,… As Table 2 shows, hedges and boosters occurred with a similar frequency (17.7% and 18%, respectively). It was found that student writers relied on hedges to express their uncertainty, exercise caution, and mitigate the statement (13 & 14). On the other hand, they used boosters such as 'must' and 'will 'to allow themselves to make claims with certainty (15).
13) In addition, it might also let people know that you have a certain level of professionalism. 14) You would better turn on the silent mode before you enter a meeting room, theatre, training class. 15) As a conclusion, hand phone etiquette is something that we must apply in our daily life. Lastly, attitude markers were the least common interactional metadiscourse used in students' essays, with only 3 cases. This may be due to the nature of the essay topics which required the students to provide more factual information, instead of emotive opinions. Therefore, they might have withheld from expressing personal feelings towards the propositions. The only attitude markers found in the corpus were the attitude verbs 'believe', 'think' and 'hope'.
16) I believe most people start up new in a college.

CONCLUSION
This study investigated the frequency and functions of metadiscourse markers used in essay writing of Malaysian L2 students. Hyland's (2005a) interpersonal model of metadiscourse was used as a framework and a top-down approach was followed to manually identify and tabulate the frequency and functions of both interactive and interactional metadiscourse markers in the context in which they were used. In summary, it was found that Malaysian L2 students used more interactive devices than interactional in their essay writing. In terms of the subcategories of interactive devices, analysis indicated that transitions and endophoric markers were the most commonly used, followed by frame markers; while code glosses and evidentials were found to be the least frequent. In relation to interactional devices, results revealed that engagement markers were the most prevalent, followed by self-mentions; whereas attitude markers were the least frequent marker of this type and hedges and boosters were used equally.
To sum up, this study demonstrated the significance of using metadiscourse markers in essay writing as a way to achieve cohesion and produce an essay of high standard and quality. The findings are beneficial for ESL/EFL students who may not be familiar with the rhetorical structure as well as the types of linguistic features used in essays in English. In addition, having an adequate knowledge of the functions of the types of interactive and interactional metadiscourse markers can help novice and struggling student writers to have a clearer picture of the organization of their write-ups, in general, and how they can project themselves in their text, engage their readers through the discourse, and evaluate their degree of commitment toward a proposition, in particular. They can use the interpersonal metadiscourse model as a reference point to organize their discourse and establish various connections with their prospective readers. Therefore, it is recommended that teachers explicitly expose students to the knowledge of metadiscourse markers and the communicative functions inherited in them in order to raise their awareness of how these markers can contribute to our understanding of cohesion and coherence in writing. Therefore, further studies can be conducted to examine the role of metadiscourse markers in other written genres relating to ESL/EFL students to accentuate their significance for second language writers. Moreover, using a larger corpus may yield more generalizable results in relation to the findings of this study.