CHINESE EFL UNDERGRADUATES’ ACADEMIC WRITING: RHETORICAL DIFFICULTIES AND SUGGESTIONS

Difficulties encountered by students in L2 academic writing has been a subject of research for several decades. However, to date, there still remains a lack of detailed and in-depth investigation into this area of interest. This qualitative study thoroughly investigated the rhetorical difficulties faced by Chinese EFL undergraduate academic writers, and collected suggestions on how to address these rhetorical issues. To be sufficiently detailed and thorough, this study divided students' difficulties into processand product-related difficulties, and used triangulated data from supervisors' perspectives, students' perspectives, and supervisors' comments to address research questions. Although there were no strong generalizations derived from data from different perspectives and sources, the findings of this study showed supervisor perceptions of the rhetorical difficulties the students experienced were almost identical. In nature these rhetorical difficulties were culturallyembedded and genre-related issues; and the degree of difficulty experienced by each student varied. In this study, supervisors and students both suggested that, to solve rhetorical difficulties, teacherstudent communication should be improved. This study provided empirical evidence to contrastive rhetoric theory and socio-cultural theory. It also offered suggestions on how to strengthen future research in this area of inquiry, and how to improve academic writing teaching in L2 educational

This study is motivated by some changes that are currently taking place in China.For years in China, academic writing in English (AWE) for undergraduates has been considered as English majors' BA thesis writing.It is generally taught in the final year of bachelor's study, and taken as the only academic writing task English majors need to fulfil before graduation.In 2006, the School of English and International Studies at Beijing Foreign Studies University (BFSU, China's top university for foreign languages studies) led curriculum reforms.This school became well aware that it is essential for underclassmen students to develop a good level of academic writing ability, and it set up an AWE course for sophomore English majors.Now, like BFSU, more schools in China are teaching or plan to teach sophomore English majors about AWE.They are coming to realize that it is even necessary to offer the AWE course to all firstyear undergraduates (both English-majors and non-English majors), since AWE involves not only English-majors' graduation thesis writing, but also the writing of some assignment tasks that undergraduate freshmen need to face.In all, these changes imply that teaching practice for AWE in China is still in its infancy, and Chinese AWE teachers need to learn how to teach more effectively.
As an understanding of the challenges students face is essential for effective teaching practice, this study thoroughly examined Chinese EFL learners' writing difficulties with AWE and collected suggestions on how to address these difficulties.This study aimed to enhance EFL, as well as ESL teachers' understanding of the difficulties and needs of L2 student academic writers in order to help L2 teachers offer better guidance to their students.
To date, there has been a number of research studies about difficulties encountered by students in AWE (e.g., Belcher & Braine, 1995;Bitchener & Basturkmen, 2006;Casanave & Hubbard, 1992;Cooley & Lewkowicz, 1997;Phakiti & Li, 2011;Qian & Krugly-Smolska, 2008;Tang, 2012;Wang & Li, 2008).The early research in this area of inquiry usually looked at the entire piece of student writing and surveyed on a large scale the student difficulties at all levels of text.Using a questionnaire, Casanave and Hubbard (1992) asked doctoral students' supervisors to evaluate their native and non-native English speaking (NS and NNS, respectively) students' writing problems.Altogether thirteen features of writing, ranging from linguistic features to features at the level of discourse, were given to the faculty to rate.This survey revealed that NNS students had greater doi: dx.doi.org/10.17509/ijal.v6i1.2645problems than NS students in "overall writing ability" (p.38), and perceptions varied between faculty members in humanities/social sciences and those in science/technology.Bitchener and Basturkmen (2006) confined their attention to student difficulties with writing the discussion of results section (DRS), and carried out a small-scale study.Using a well-designed schedule to interview four supervisor-student pairs, the researchers found there was a limited common understanding between supervisors and students.This study showed that the supervisors saw students' limited understanding of the DRS as a genre as the source of problems, while the students simply attributed all their difficulties to a lack of language proficiency.Moreover, this study revealed some causes of the students' difficulties, which included students' inability to make generalizations, negative transfer of their native language, and their superficial reading of literature.Generally speaking, the results of this study revealingly captured the necessity to help students understand the rhetorical genres of different parts of a thesis and pointed out the future directions this area of inquiry needed to take.
Qian and Krugly-Smolska ( 2008) approached questions quite similar to those asked by Bitchener and Basturkmen (2006), but the former focused on difficulties with writing a literature review (LR) and did not adopt a mode that grouped supervisors and students as pairs.Only using student interviews, Qian and Krugly-Smolska (2008) collected opinions from four Chinese master's students who were studying in Canada and who had varied previous experience of LR writing.Their study yielded findings similar to Bitchener and Basturkmen (2006).In this study, students also considered language issues (mainly the problems in vocabulary, sentences, and paraphrasing a section from a source) constituted their major obstacle.It is not clear whether triangulating the interview data from students with other types of data (e.g., data from teachers or other sources) could offer more insights or not.
Also using interviews, Wang and Li (2008) identified the challenges international ESL doctoral and master's students experienced in their thesis writing process and sought to understand needs of these students.The researchers found the challenges the students encountered included language problems (e.g., the difficulty in expressing themselves clearly and accurately) and the negative influence of their cultural background on their thesis writing.What is interesting is that, although feeling inadequate and unconfident in using English while writing, the students in this study felt dissatisfied with receiving feedback focused only on language.However, in this study, Wang and Li failed to go into details about the findings their study produced.
From a brief overview of writing difficulty studies, it can be seen that these studies placed great emphasis on the experiences of ESL postgraduates.Increasing research attention being paid to this group of students can be explained by the fact that the number of international ESL students seeking postgraduate study in English-speaking countries has grown very fast over the past several decades.In fact, it is equally important to know about their experiences of doing AWE tasks in their home countries as EFL undergraduate learners, and it is worthwhile to give careful attention to students studying at undergraduate level in the EFL academic contexts.
At the same time, a review of previous literature shows that there is a lack of detailed and in-depth investigation along this line of inquiry.Previous studies examined a wide range of issues and challenges L2 academic writers encountered.This large research scope prevents researchers from going beyond "a mere identification of writing difficulties" (Bitchener & Basturkmen, 2006, p. 14) and revealing the root causes and nature of these difficulties.Now, researchers need to conduct more detailed studies so as to offer a micro-analysis of student difficulties and give further substance to the existing findings.
Generally, it is believed that ESL/EFL academic writers usually face two essential issues: syntax (language) difficulties and rhetorical difficulties (Kroll, 1990;Reid, 2006).When it comes to future difficulty studies, Bitchener and Basturkmen (2006) suggested that specific difficulties related to language should be categorized separately from those difficulties at the rhetorical genre level.Considering that "foreign students who have mastered syntactic structures have still demonstrated inability to compose adequate themes, term papers, theses, and dissertations" (Kaplan, 1966, p. 13), and that few studies in L2 scholarship deal with rhetorical concerns (Tardy, 2005), this study chose to focus on the rhetorical difficulties of L2 students in academic writing.
Before moving on, here it is important to be specific about what difficulty in rhetoric means in this study first.This study followed a commonlyused conceptualization of rhetoric in the literature of applied linguistics, and defined it as discourse-level organizational patterns (Kaplan, 1966;Casanave, 2004).More specifically, it was composed of the following elements: 1) limiting and focusing on the topic in a manner appropriate to its overall approach and length, 2) remaining focused on the topic throughout, 3) creating and using paragraphs effectively, 4) maintaining a consistent point of view, 5) sequencing ideas in a logical manner, and 6) using coherence and cohesion devices appropriately and as necessary (Kroll, 1990).To be sufficiently detailed and thorough, difficulty in this study was precisely classified into two types: process-related difficulties and product-related difficulties.Processrelated difficulties represented the difficulties students experienced during the process of academic writing, and product-related difficulties represented the problems/errors that appeared in the final written product as symptomatic of student difficulties.In brief, rhetorical difficulty in this study was taken to mean process-and product-related difficulties in terms of the six specific rhetorical elements.
In summary, given the recent curriculum reforms in China, the necessity to understand challenges Chinese EFL students face, and the research gap that studies at undergraduate level are comparatively rare, this study gave its attention to Chinese EFL undergraduate students.To allow more in-depth investigation of student difficulty, this study focused on rhetorical challenges and precisely defined key terms.Specifically, in this study, two main concerns were addressed: 1.While Chinese EFL university student writers are dealing with AWE tasks, is each of these six rhetorical issues difficult for them to handle?If so, what is hard and why is it hard?Is it still a problem in the finished writing, and what leads to the problems?2. What is the best way to cope with each of these rhetorical difficulties?

METHOD
In the case when a problem or issue needs to be thoroughly explored, it is believed that qualitative research is very useful (Creswell, 1998;Gall, Gall, & Borg, 2003;McKay, 2006;Merriam, 1988;Yin, 2009).As such, this study, which aimed to implement a detailed and in-depth study, took a qualitative study approach.Specifically, this study used interview data from the perspectives of supervisors and students, and document data from supervisor comments on different drafts of student writing to address research questions.

Participants
As noted at the outset, in many universities in China, the only piece of academic writing done in English during students' undergraduate degrees is the finalyear BA thesis that English majors compose.Thus, for ease of data collection, senior students and their thesis supervisors (that is, the teachers who provided advice on graduation thesis writing) from an English Studies Department at a university in Beijing, China, were invited to participate in this study.The student participants were chosen according to one criterion.That is, the topics of their BA theses should cover different research areas of English majors (e.g., literature, translation, culture studies, linguistics) in order to see whether their difficulties were shared across research areas.Three students, Gill, Jane, and Sherry (pseudonyms), whose theses were about medio-translatology, literature (contemporary British novel), and American culture and society accepted the invitation and generously agreed to share all their drafts, including supervisor comments on their drafts, with the researchers.The study was then explained to the students' thesis supervisors: Erin, Jewel, and Lynn (pseudonyms).They were willing to take part in an interview and gave the researchers permission to utilize all comments they had provided on students' drafts.Table 1 provides additional information about each supervisorstudent pair.In this study, pseudonyms were used for participants to protect their anonymity.Thesis writing-related context information BA thesis writing was a task these student participants undertook in the final semester (semester eight) of their college study.In semester seven, an AWE course was set up to prepare them for thesis writing.The course was a nine-week credit-bearing compulsory course with the focus being to develop academic writing competence of English-major seniors.The teacher of this course, who happened to participate in this study (Erin), was a professor of British and American literature.Erin believed that during the course, she had explained clearly to the students how to handle the AWE task, although she usually used literature thesis writing as examples.The students also received sample theses from the department.They were literature theses as well.Also, it is worth noting that in semester six the student participants had ever been asked to write course papers for earning 25-30 percentage of overall course grade.However, none of them considered these tasks to be their "genuine" AWE experiences.This was because they felt they knew nothing about AWE at that time and they performed these writing tasks in the same way they wrote informal essays.It can be said that the three student participants had never dealt with "real" AWE writing until they wrote their final thesis.

Data collection
As mentioned above, two types of data were used in this study: interview data and document data.To collect interview data, semi-structured interviews were employed because this approach has the advantage of providing reasonably structured data, but with greater depth (Gall, Gall, & Borg, 2003).The interviews with supervisors were organized around three topics: (1) general description of your experience of supervising the student participant, (2) perceptions of major student difficulties and problems, and (3) perceptions of student difficulties and problems with the specific rhetorical elements listed by Kroll (1990), and suggestions for each particular difficulty.The interview protocols for students were modified on the basis of the prompts for supervisors.
During the course of interviews, the interviewers/ researchers followed Richards' (as cited in Bitchener & Basturkmen, 2006) recommendation and avoided sticking rigidly to the interview schedule so as to provide the interviewees with sufficient "thinking space" and encourage them to offer extensive responses.All the interviews were conducted primarily in Chinese according to interviewee preference.Interviewing in Chinese also ensured the interviewees' accurate articulation of perceptions.Each interview lasted about 50 minutes, and was audio-recorded.All participants were interviewed once and separately.
Document data in this study consisted of supervisor comments.At the end of each student interview, the student participants were asked to email to the researchers their thesis drafts and share the comments their supervisors had written on their drafts.However, the quantity of the drafts varied for each student: Jane had two commented drafts; Gill and Sherry had three.

Data analysis
In this study, interview data were processed by a qualitative content analysis.First, based on the research questions, the analytic categories were identified.They were 1) process-related rhetorical difficulties students encountered during thesis writing process, 2) product-related rhetorical problems that appeared in students' finished thesis, 3) hard parts of each difficulty, 4) reasons for each difficulty, and 5) suggestions to improve in terms of each difficulty.Then, the interview recordings were transcribed, segmented, and coded.The data were transcribed according to Bitchener and Basturkmen's (2006) method of transcription.That is, actual words and pauses were transcribed, but intonation, non-verbal cues, and other phonological details were not.The participants checked the accuracy of a portion of the transcribed data and verified its accuracy.Then, the transcribed data were coded, and then the coded data were sorted into the categories that had been established.Document data (teacher comments) were used to identify the two types of students' rhetorical difficulties.The negative comments associated with the six specific rhetorical elements were sorted out first.If one of them was repeatedly found within one draft or across drafts, it meant the student had difficulty with the corresponding rhetorical element during thesis writing and it was considered as a process-related difficulty.If the same (or a similar) comment persisted in students' final drafts, it was referred to as a product-related difficulty.

FINDINGS
To address research questions, this study used data from different perspectives and sources.In what follows, the findings from thesis supervisors' perspectives, students' perspectives, and thesis supervisors' comments are reported in three separate parts.These three parts all are related to the two main concerns of this study, rhetorical difficulties and suggestions.

Process-related rhetorical difficulties and their causes
In this study, supervisor perceptions of the rhetorical difficulties students experienced during the process of their thesis writing were almost identical.The process-related difficulties the three supervisors all identified were 1) limiting the topic, 2) creating effective paragraphs, 3) creating a logical sequence of ideas, and 4) appropriately using coherence devices.Lynn's perception slightly differed from Erin's and Jewel's; she pointed out that her student (Sherry) had an additional difficulty with remaining focused on the topic throughout the thesis.
Moreover, the participating supervisors offered multiple insights into what was hard for their students and why these difficulties were hard for them.The supervisors agreed that, because their students were weak in critical thinking ability, they had difficulty in forming a critical analysis of the previous studies, going from previous studies to their own, and limiting their topic when writing a literature review.Also, Erin, Jewel, and Lynn said that their students seemed to "lack ability to make generalizations", which made it difficult for them to extract key points; and then they were unable to write good topic sentences, and limit the writing topic.Besides these, according to the supervisors, differences in the Chinese and English way of thinking (that is, Chinese spiral thinking mode and English linear thinking mode) and writing (that is, Chinese roundabout way of writing and English direct way of writing) caused the student difficulty in writing effective paragraphs.They believed that, under the influence of the Chinese way of thinking and writing, their students usually had difficulties in expressing an idea straightforwardly at the beginning of a paragraph by using a topic sentence and directly developing this idea within the paragraph.What is more, the supervisors felt the students had no idea that it was necessary to write introductory and concluding remarks in different chapters, and they did not know how to write these remarks.The supervisors considered that these difficulties resulted from a lack of logical thinking, as well as from the differences between Chinese and English way of thinking and writing.
Among all the reasons given, there were two other causes of difficulty the supervisors frequently referred to.One reason was that the students' limited knowledge and understanding of the content and structure of AWE contributed to their difficulties.The other reason was related to supervisor-student communication.Erin and Lynn often noted the students had difficulty in understanding what they meant or what their comments meant.Erin said, "Before thesis writing, I clearly explained once again to her (Gill) how to write from subject to topic, and how to write from topic to thesis statement.But it seemed that she didn't listen to me or she didn't understand and internalize what I had said to her."

Product-related rhetorical difficulties and their causes
With regard to the product-related rhetorical difficulties appearing in the finished thesis, both Erin and Lynn perceived that their students' process-related difficulties persisted.They felt these difficulties stemmed from their students' low level of logical thinking ability and having no idea regarding what content should be included in different thesis sections and how to organize these sections.Jewel, however, felt that her student (Jane) overcame her process-related rhetorical difficulties, and her final thesis draft was not problematic rhetorically.Jewel reasoned that this was likely due to "the easy teacher-student communication, and Jane's quick understanding of teacher comments and teacher instructions."In brief, the supervisors' perceptions of the students' difficulties in rhetoric are summarized in Table 2 according to the six rhetorical aspects.Jewel talked about the importance of written communication, teacher written feedback.She considered that students usually could not discover the rhetorical problems existing in their writing on their own, and teachers needed to make full use of written comments to to help the students with rhetorical issues.In addition, Erin, Jewel and Lynn all referred to classroom instruction.They thought the basic writing course should fully prepare the students to be able to develop effective paragraphs, produce well-arranged essays, and use linking words appropriately.With respect to academic writing instruction, in Jewel's words, "the teacher could lead the students to closely read well-written academic papers in order to help them develop a sense of how to make a critical analysis of previous literature, how to connect chapters with introductory and concluding remarks, and how to organize academic papers." All supervisors also believed that teachers needed to know well their students' subject of study and the literature related to it."Otherwise", Erin said, "teachers could not offer help when their students were unable to argue, or strongly argue for their ideas, and when they have difficulties in creating paragraphs effectively."

Process-related rhetorical difficulties and their causes
During the interview, Jane said her thesis-writing experience was "painful".First, Jane felt it was difficult to remain on the topic.She said, "since my thesis involved too many novels, while writing, I needed to keep going back to read the paragraphs that had been finished and making frequent revisions in order not to be off topic."Another issue Jane found troublesome was developing effective paragraphs.She said this was because "it's difficult to find evidence or the most persuasive evidence to support the topic sentence."The third rhetorical issue that was hard for Jane was to maintain a consistent point of view.According to Jane, "this difficulty arose because I changed my viewpoint as writing continued and I read more literature.I had to take time and effort to reorganize my ideas, or even readjust the organization of my thesis that had already been well designed before thesis writing."What is more, Jane felt she met difficulties in logically sequencing her ideas.While writing, she said she had to "keep changing the sequences of paragraphs to ensure a logical sequence."Jane believed this difficulty was also caused by the occurrence of new and better ideas in her mind as her writing progressed.
Gill and Sherry did not report many processrelated difficulties; however, Sherry noted that she had difficulty in "remaining on the topic throughout the thesis."She thought the discussion chapter in her thesis was off topic and should not be included in her thesis.She said her difficulty lay in that she had no idea why it needed to be included in her thesis, and she had nothing to say when writing that chapter.She felt she just wrote this chapter according to her supervisor's requirements.

Product-related rhetorical difficulties and their causes
Jane thought she had three product-related difficulties.First, she believed that the topic of her finished thesis was too broad since she failed to define a key construct in her thesis.Second, Jane believed that several paragraphs in the fourth chapter of her thesis were irrelevant to her topic.She said she wrote these paragraphs just for more words and they were "meaningless and off her topic."The third problem Jane felt she had was "creating effective paragraphs."Jane said, "In my thesis, some of my opinions and their supporting evidence were put in two paragraphs."She felt that this problem occurred because of lack of clear thinking.
Still, Gill and Sherry did not report many product-related difficulties.Sherry insisted that her discussion chapter was "full of empty words," and her process-related difficulty turned out to be a weakness of her finished thesis and a productrelated difficulty.Gill also identified one productrelated rhetorical difficulty, that is, the paragraphs in the main body of her thesis were not effective enough.She considered that the reason for this was that there were too many culture-related issues to cover in her thesis, and consequently her paragraphs lacked depth.Table 3 summarizes the students' perceptions of their own difficulties in rhetoric according to the six rhetorical aspects.Note: D represents process-related difficulties.P represents product-related difficulties/problems.+ indicates the presence of difficulty.The numbers 1, 2, 3, 4, 5, and 6 represented the six rhetorical aspects.

Students' suggestions
The suggestions Gill, Jane, and Sherry put forward were not very specific, but all were related to teacher-student communication.Gill considered that it was quite necessary for the students to follow their thesis supervisors' instructions to work out a wellplanned outline before thesis writing and that supervisors' quick response on each draft was crucial.Besides devising an elaborate outline and identifying the key words of the thesis, Jane felt that it was important for teachers to focus on structure and logic when providing feedback.Otherwise, she said, "students would devote their attention only to language errors."Sherry suggested that the students should clearly express their opinions to their supervisors; she felt her thesis writing was rather difficult because she dared not tell Lynn that she thought the discussion chapter was irrelevant to her topic.

Rhetorical difficulties: Findings related to supervisors' comments
In general, the supervisors did not provide many comments on rhetoric but mainly on what to be included in the thesis and how to write theses in a more broad sense.Within one draft and across Gill's drafts, Erin continuously pointed out that Gill made no critical analysis of previous literature; her supporting evidence was insufficient; there was excessive use of long quotations without illustration, and introductory and concluding remarks were needed between chapters.On draft three, Erin pointed out these problems once again.Erin's comments indicated that Gill had process-and product-related rhetorical difficulties in limiting and focusing on the topic, creating effective paragraphs, and using coherence devices appropriately.Jane had only two drafts commented by Jewel.In her comments on draft one, Jewel frequently mentioned insufficient explanation of topic sentences and lack of introductory and concluding remarks.On draft two, Jewel wrote comments only on language and mechanics.Jewel's comments showed that Jane had difficulties creating effective paragraphs and using coherence devices.However, Jane overcame these difficulties and they did not end up as her product-related difficulties.
The comments Lynn made on Sherry's three drafts centred on what content should be included in the thesis and how to organize theses.Thus, concerning the specific rhetorical issues, Lynn only mentioned Sherry had problems with using cohesive devices and writing introductory and concluding remarks.The rhetorical difficulties derived from the supervisors' comments are summarized in Table 4 according to each type of rhetorical difficulty.Note: D represents process-related difficulties.P represents product-related difficulties/problems.+ indicates the presence of difficulty.The numbers 1, 2, 3, 4, 5, and 6 represented the six rhetorical aspects.

DISCUSSION
This section contains two main parts.Findings about the two main concerns of this study, rhetorical difficulties and suggestions, are discussed in turn.

From supervisors
In this study, the supervisors reached agreement and identified common process-related difficulties.However, Jewel pointed out that the "processrelated difficulties" the supervisors all referred to did not turn out to be Jane's "product-related difficulties", while Erin and Lynn considered Gill and Sherry's "process-related difficulties" persisted and ended up as their "product-related difficulties."This difference shows that there is variance in the degree of difficulty students experience.More importantly, this result indicates that classifying "writing difficulties" into types offers a new understanding of student difficulties since it is revealing about the extent to which students experience difficulties.Certainly, defining "writing difficulties" by classifying them in this way is still very general.It is strongly suggested that the construct "writing difficulties" be more scientifically defined in future studies (Bitchener & Basturkmen, 2006).
To a great extent, the varying degrees of difficulty mentioned above may be due to students' different ability levels, because Jane was considered to be a student with high writing and overall ability.However, in her interview, Jane mentioned that she viewed writing BA thesis as an opportunity to prepare academically for her master's study in the UK, which was to begin three months after she received her BA degree in China.This suggests the variance in difficulty levels could also be explained by students' strong motivation to write their thesis well.According to the findings of previous studies (e.g., Campbell & Li, 2008;Phakiti & Li, 2011), students' motivation, self-efficacy, positive attitude towards difficulty, and academic English proficiency may help to ease their academic difficulties.However, when it comes to academic writing difficulties, factors leading to the reduction of degree of student difficulties warrant further study.
Supervisors in this study commonly expressed that problems in critical and logical thinking contributed to process-, as well as product-related difficulties.Previous studies suggested that Chinese students had problems and difficulties in thinking critically and logically, and the reasons were that little emphasis was put on cultivating students' critical and logical thinking ability at schools in China (Tian & Low, 2012).However, to a large extent, it is undeniable that Chinese culture, which favours harmony and depreciates external criticism (Taylor & Chen, 1991), prevents students in China from being critical or expressing criticism.Besides, according to Leki (1992), what an argument is, what constitutes proof of an argument, what is relevant or irrelevant, and what is logical or illogical all are culturally determined.In this sense, the nature of writing difficulties related to critical and logical thinking are culturally embedded challenges.
In this study, the supervisors also often referred to differences between English and Chinese ways of writing and thinking to explain the reasons for the student difficulties with topic sentence writing, cohesion, and the like.This finding, likewise, supports the view that the rhetorical difficulties Chinese EFL students face are cultural challenges.This point has been corroborated by Wang and Li's (2008) study, which revealed that Asian cultures and language are used to a writing convention that values a delay of the central argument towards the end of a paragraph instead of putting it straightforward at the beginning.In fact, from the perspective of contrastive rhetoric theorists, who claim that cultural values underlie writing in different languages, the view that rhetorical difficulties are cultural issues can also be confirmed.
At the same time, the process-related rhetorical difficulties Chinese EFL undergraduates experienced and product-related difficulties that resulted are a genre-related challenge.This generalization is supported by the consensus reached among the participating supervisors that the students' knowledge and understanding of the content and structure of AWE were limited.This perception is also evident in Bitchener and Basturkmen's (2006) study, where supervisors considered that students experienced a high level of difficulty in meeting the requirements of genres while writing DRS.Additionally, there are several other studies which have indicated that Chinese undergraduates need to be explicitly exposed to the different genres of academic writing (e.g., Qian & Krugly-Smolska, 2008;Wang & Yang, 2012).
Generally, the reasons for the two types of rhetorical difficulty that supervisors identified in this study can be classified into two categories: 1) reasons related to the students' abilities and knowledge, and 2) reasons related to supervisorstudent communication.As the reasons related to student attributes are consistent with the results of many previous studies, it is not surprising that supervisors in this study provided the causes such as students' inability to make generalizations (Bitchener & Basturkmen, 2006), a negative transfer of students' native language (Bitchener & Basturkmen, 2006;Dong, 1998;Kaplan, 1966;Wang & Li, 2008), and so on.To some extent, it is not anticipated that teacher-student communication was uniformly perceived by the supervisors as the other reasons.This shared perception also reinforces the views that effective collaboration and communication between supervisors and students are crucial for helping the students develop academic literacy competence (Belcher, 1994;Dong, 1998;Wang & Yang, 2012).

From students
In James' (1984) study, L2 doctoral students unanimously asserted that they found developing ideas to be challenging.However, in the present study, the students hardly concurred with each other and identified no common difficulties.This might result from the students' unfamiliarity with the genre of AWE in general and with the particular genre conventions of their research areas.As mentioned in Methodology section, before writing theses, the students had never dealt with "real" academic writing in English.At the same time, the academic writing class the students attended was somewhat oriented to writing a literature thesis.Lack of practice and exposure to AWE genre requirements definitely leads to the students' low familiarity with the AWE genre and influences their perceptions and explanations of the difficulties they experienced and faced.
As a matter of fact, the reasons the students provided in this study to explain their difficulties, such as having nothing to say or finding no evidence to support the topic sentence, also indicate that the participating students' understanding of the AWE genre was limited.To a large extent, it can be said the students failed to identify the precise underlying reasons for their process-and product-related rhetorical difficulties.For example, Gill considered that some paragraphs in her thesis were superficial and ineffective, but this was not because, as she had said, there were too many cultural issues her thesis needed to explain.In large part, it was because her topic was still too broad and needed to be further narrowed down.In general, as student difficulties can be explained by their unfamiliarity with genre requirements of academic writing, it seems valid to say students' difficulty in rhetoric is basically a genre-related problem.

From supervisors' comments
Considering the supervisors shared many opinions about student difficulties in their interviews, it is expected that there be a similar concurrence of ideas present in their comments.However, according to the supervisors' comments, there was only one common rhetorical difficulty identified, that is, appropriately using cohesion and coherence devices.On the whole, the common student difficulty and problem related to cohesion and coherence, which included unclearly signposting connections between sentences, paragraphs, and chapters through inappropriately using linking words or introductory and concluding remarks, can be generalized as an issue with the use of metadiscourse (that is, an array of devices in a written text that is used to connect ideas, and signal sequences of topics).This perhaps can be explained by the fact that Chinese writing favours simplicity (Hinds, 1990) and generally uses fewer metadiscourse features than English writing (Kim & Lim, 2013).From this perspective, the students' difficulty in rhetoric is a culture-based challenge in essence.

Suggestions from supervisors and students
Despite the range of differing opinions concerning student difficulties, the supervisors and students in this study all referred to teacher-student communication when asked to provide suggestions for dealing with rhetorical difficulties.Suggestions such as frequent supervisor-student communication, student willingness to communicate with the supervisor, face-to-face feedback, and feedback focused on rhetorical issues, once again support the view that AWE at university is not an isolated, independent task, but a process of composing and learning through interaction and collaboration.Wang and Yang (2012)'s study indicated that, to establish good supervisor-student communication, students needed to make sufficient preparation for meetings with supervisors, actively participate in negotiating with their supervisors, and take a positive attitude towards supervisor-student negotiation.In fact, to communicate effectively, it is vital that both sides take an active role.

CONCLUSION
Using interview data and document data, this qualitative study investigated the rhetorical difficulties that hindered Chinese EFL undergraduates during academic writing process and the problems that appeared in students' final written production as symptomatic of their difficulties.This study also attempted to uncover corresponding solutions.Although there were no strong generalizations regarding challenges commonly faced by EFL student academic writers derived from this study, it is possible that these rhetorical difficulties are culturally-embedded and genrerelated in nature and that the degree of difficulty experienced by each student varies.To solve these difficulties, both the supervisors and the students suggested that teacher-student communication should be improved.These findings provide empirical evidence to contrastive rhetoric theory that each language and culture has some rhetorical conventions it prefers, and to socio-cultural theory that interaction and collaboration play a key role in L2 writing development (Storch, 2013).
In addition, the findings of this study have implications for considering future research designs.First, since it was difficult to make generalizations of findings across the perspectives and angles in this study, an improvement for future research design would be to ensure homogeneity in participants in terms of research field, writing and overall ability, and so on.Additionally, some other methodological issues, such as when to conduct interviews and how to word interview questions, warrant particular attention.Specifically, future research should consider timing as an issue; for example, conducting a similar study several weeks after thesis completion would be problematic, as student recall of the struggles and frustrations experienced during thesis writing may be less accurate.The meaning of the term "writing process" should be made clear in interview questions, as it might be interpreted to refer to either the draft-writing process itself, or to the cyclical process during which students write drafts, teachers provide comments, and students use teacher comments to revise drafts.
The findings of this study offer several implications for teaching English writing in both Chinese EFL and international ESL contexts.First, according to socioculturalists, the teacher's expert instruction should be structured around "what a learner is currently able to do alone", or a "learner's 'readiness'" (Gibbons, 2002, p. 10).That is to say, L2 writing teachers need to keep tailoring their classroom instruction and different types of feedback according to the students' cognitive levels, proficiency levels, as well as affective needs in order that their students can understand what they really mean and communicate with them well.Second, considering students have insufficient knowledge of academic writing, it is necessary for the students to be exposed explicitly to the AWE conventions in general and specific to their own research area prior to academic writing.Additionally, in preparation for AWE, it is necessary for teachers to involve the students in truly experiencing the English way of presenting thesis statements, writing topic sentences, developing ideas, connecting and sequencing ideas, and so on, in order to cultivate students' critical thinking ability and logical thinking ability.Equally important, EFL, as well as ESL writing teachers, need to develop cultural sensitivity in their teaching practices so as to better assist the academic writers from the Chinese culture to respond to those rhetorical challenges aroused by culture differences, such as the use of metadiscourse (e.g., use of signposts that indicate how a chapter/section is organized), critical argumentation, writing of topic sentences, and so forth.Last but not least, L2 writing teachers also need to help their students adopt a positive attitude towards AWE and take student motivation for AWE writing into account.

Table 2 .
Supervisors' perceptions of student process-and product-related difficulties across six rhetorical aspects

Table 3 .
Students' perceptions of their processand product-related difficulties across six rhetorical aspects

Table 4 .
Supervisor comments on student process-and product-related difficulties across six rhetorical aspects