Generic Complexity in Bachelor’s Theses by Chinese English Majors: An SFL Perspective

For Chinese English majors, bachelor’s theses are regarded as the most substantial piece of writing in their undergraduate study. However, in the field of EAP/ESP and genre research, thesis writing at the undergraduate level has been relatively neglected. From the perspective of genre in Systemic Functional Linguistics, the macrostructure of a bachelor’s thesis constitutes a macrogenre which combines more than one elemental genre to accomplish complex goals. An understanding of the deployment of elemental genres in this macrogenre can further help reveal the rhetorical values circulating in this thesis writing community. Based on the genre taxonomy developed by SFL genre theorists (e.g., Rose, 2010, 2015a, 2015b), this study analysed the genre deployment of 40 highly-rated bachelor’s theses written by English majors at a Chinese university, triangulated with semi-structured interviews with thesis writers and advisors. The results show that the 40 theses contained 776 shorter texts instantiating 22 types of elemental genres across 7 genre families. Specifically, reports were most extensively used by the thesis writers to transmit their received disciplinary knowledge. Arguments and text responses, though ranking lower in number, were essential to the writers’ projection of evaluative meanings and authorial selves. Stories, chronicles, explanations, and procedural genres were deployed sporadically, but empowered the writers to perform a constellation of social roles. The findings of this study may assist novice thesis writers by heightening their genre awareness, and more practically, increasing their knowledge on the specific types of elemental genres over which they need control to produce a rhetorically well-developed bachelor’s thesis. This paper then concludes with implications for teaching and researching thesis writing in non-English dominant contexts.


INTRODUCTION
In the field of EAP/ESP and genre research, compared with the higher-level academic genres, such as postgraduate theses or dissertations, published research articles, or textbooks, theses at the undergraduate level have been relatively neglected. One arguable reason for the lack of attention comes from the (mis)belief that bachelor students do not so often create new knowledge as the advanced writers do, but only transmit received wisdom (Grobman & Kinkead, 2010;Xu et al., 2016). A few studies have examined the textual features of bachelor's theses, or equivalent genres, across disciplines (Gardner & Holmes, 2009;Nesi & Gardner, 2012;Hyland, 2012). For example, Hyland (2012) examined the use of stance and voice markers in 64 project reports written by final-year Hong Kong undergraduate students from eight disciplinary fields, in comparison with a reference corpus of research articles written by 'expert writers' from closely related disciplines. Most of the previous studies, English (i.e., the configuration of elemental genres in the macrogenre), and behind that, the underlying rhetorical values adopted by this writing community. The present study, then, is an attempt to bridge this gap. Before diving into the methodological waters, we shall first lay out the theoretical framework for this paper, namely the SFL's view of genre and its taxonomy of key written genres.

SFL'S VIEW TOWARDS LANGUAGE AND SOCIAL CONTEXT AND THE DEFINITION OF GENRE
At the heart of SFL is an idea of language as semiotic resources for meaning making in social contexts (Halliday, 1978). Halliday argued that the way into an understanding about language lies in the study of texts in their social contexts and then he described social context as 'the total environment in which a text unfolds', which is stratified into two levels -'context of situation' and 'context of culture' (Halliday, 1978).
Context of situation constitutes the register of a text and varies on three general dimensions, i.e., field, tenor, and mode (Halliday & Matthiessen, 2014). J.R. Martin noted, however, that the three register variables cannot be associated with the social purpose of texts and thus proposed the concept of genre as a variable lying in the context of culture, which is reflected in the structural patterns through which a text unfolds to achieve its social purpose (Martin, 1992;Martin & Rose, 2008;.
Genres are thus defined in SFL as staged, goal-oriented, social processes (Martin, 1997;Martin & Rose, 2008;Rose, 2015aRose, , 2015b: social since texts are always interactive events and wedded in the social context; goal oriented in that a text unfolds towards its interactants' purposes; staged, because it usually takes more than one step to accomplish the goal.

TAXONOMY OF KEY EDUCATIONAL GENRES AND GENRE FAMILIES
Research endeavours in this approach, led by J.R. Martin and based mainly on a series of large-scale action research across the primary and secondary school curricula, have richly identified a broad range of key written genres through which knowledge is transmitted. A global sketch of these genres is presented in Figure 1. Here, genres are grouped into seven genre families functioning, to varying degrees, on a continuum of three most general social purposes: (1) to engage and entertain through stories and historical chronicles; (2) to inform, i.e., to provide information, through some factual genre families, such as, explanations, reports, and procedural genres; and (3) to evaluate opinions/issues via arguments or texts via text responses. Detailed descriptions of each of these genres, in terms of their primary social purposes and common stages and phrases in particular, have been richly documented in the works of Martin (1997), Rothery & Stenglin (1997, Veel (1998), Coffin (2006), Martin and Rose (2008), Rose (2015aRose ( , 2015bRose ( , 2017, and others. Here, efforts to reiterate shall be spared.

To engage & entertain
To inform GEMA Online ® Journal of Language Studies Volume 19(4), November 2019 http: //doi.org/10.17576/gema-2019-1904-16 eISSN: 2550-2131 ISSN: 1675-8021 308 It is worth pointing out that the primary and secondary school texts that the functional linguists have been developing their genre theories around were mostly short and had 'relatively simple purposes' (Martin & Rose, 2012, p. 1) -thus glossed as 'elemental genres'. In recent years, genre research in this school has expanded beyond into tertiary education and workplaces (e.g., Jordens at al. 2001;Jordens & Little, 2004;Hood, 2010;Humphrey et al., 2010;Nesi & Gardner, 2012;Tribble & Wingate, 2013;Coffin & Donohue, 2014;Dreyfus et al., 2015;Humphrey & Macnaught, 2016;Szenes, 2017), and it has been argued that students need to move from controlling elemental genres to writing longer, more sophisticated texts that draw on various genres to reach more complex goals as they advance on academic or professional ladders (Szenes, 2017).
Additionally, as a concept located in the context of culture, genres are defined as a recurrent configuration of meanings and these recurrent configurations enact the social practices of a given culture (Martin & Rose, 2008;Rose, 2010). Similarly, in the macrogenre which shapes and is shaped by its community of practice, the recurrent configurations of elemental genres, that is, how the generic resources are deployed and distributed, enact and manifest the rhetorical values circulating in and defining this community in focus. For novice writers who are on the threshold of seeking a membership into this community, their chances of success are largely dependent on whether or not they can write in a way that aligns with the established norms of the target community. For this reason, we are convinced that acquaintances with and concomitant efforts to observe the community's rhetorical values are of paramount importance.
Framed within these SFL-informed genre theories, the present study aims to answer the following two questions: 1. What elemental genres are used by Chinese English majors in composing bachelor's theses, and in which way are these elemental genres deployed? 2. What are the rhetorical values circulating in the thesis writing community related to the deployment of elemental genres?

METHODOLOGY RESEARCH CONTEXT
This study was carried out in the English Department in a public university in the Southwestern China. Students in this Department receive training in 2 general English writing courses in the second year of their study and practice writing skills across several rhetorical modes, including narration, description, exposition, and argumentation. In the sixth semester, the students take a 10-week course of Academic Writing, during which they are introduced to the MLA writing conventions and the bachelor's thesis requirements for English majors. According to the Writing Norms of Bachelor's Thesis for English Majors issued by the Department, English majors' bachelor's theses must be written in English only, at a minimum length of 4000 words, conforming to a rigid format that encompasses the following elements: namely, a cover page with thesis title and author/supervisor information, a contents page, an abstract and keywords in both English and Chinese, the body of the thesis, bibliography, acknowledgements, and appendices (if any). At the beginning of the seventh semester, each student is assigned to a thesis advisor, and both parties work together through a number of conferences, drafts, and revisions, until the thesis is finalised in the eighth semester. All the thesis writers have to go through an oral defense. Ultimately, 50% of the final score for each thesis is given by the thesis advisor and the other 50% comes from the committee sitting in the defense, both of which are based on an assessment rubric formulated in Chinese by the Department.

DATA COLLECTION
To analyse the genre deployment, we selected 40 bachelor's theses produced by English major graduates from this Department from the past five years. Out of a full mark of 100, 85 was set as the cut-off point, because, from informal talks with the Department directors and faculty, theses scoring 85 and above were highly regarded by thesis advisors and examiners, reflecting, in our assumption, the preferred generic patterning of this macrogenre. Out of a pool of 336 theses produced between 2014 and 2018, 63 (18.8%) met the criterion, among which 24 were written on subject matters in translation/interpreting studies, 17 in linguistics/applied linguistics, 16 in cultural studies, and 6 in literary studies. We then selected 40 theses by adopting quota sampling techniques. Ultimately, the breakdown of theses across the four subfields was 16, 13, 9 and 2, almost in the same proportion as in the entire population, with an approximately equal number of theses from each year. On all accounts, every effort was made to warrant the representativeness of the final corpus. As a final step, the selected theses were referred back to the thesis advisors to confirm them as legitimate exemplars without using any illegal tactics.
Since bachelor's theses are multimodal in most cases, a decision was made to discard the visual elements in the sampled theses, such as images, diagrams, figures, and tables, together with auxiliary texts such as the cover page, abstract, acknowledgement, bibliography, and appendices (if any). That is, only the essential verbal texts were included in the final corpus. The texts were then numbered by 'Year_No.' for subsequent analysis and reference, e.g., Thesis 2018_1, 2017_5, 2015_7, and so forth.
To supplement and triangulate the textual data, another two types of qualitative data were collected: documents and interviews. To understand the institutional requirements, two documents related to bachelor's thesis writing were examined, i.e., Writing Norms of Bachelor's Thesis for English Majors and Assessment Rubric. In addition, we interviewed 4 thesis advisors and 7 thesis writers in order to draw their insider perspectives. The 4 thesis advisors were selected on the basis of their extensive experiences in supervising and availability at the time of the study. Meanwhile, based on the writers' availability, we approached the 8 thesis writers who just defended their theses in 2018 and ultimately 7 responded to our request. Interviewees were approached with informed consensus, and all the interviews were conducted in Chinese and audio-recorded.
The interview questions fell into two categories. The first was in a semi-structured format, containing open-ended questions over which the interviewees were encouraged to speak freely about their experiences in supervising/writing the bachelor's thesis, their perceptions over its generic structures, the values that they placed on the varying genres employed, and their interpretations of and (sub)conscious conformity with the rhetorical principles imposed by the relevant documents. The second were 'talk-around-texts' questions (Lillis, 2008) which were formulated based on our analysis of the corpus. Appendix A includes the planned interview protocols with the actual questions being asked.
These qualitative data, when integrated with textual evidences from the corpus, were intended as useful sources of information for understanding the rhetorical values in the thesis writing community.

DATA ANALYSIS
Genre analysis of the corpus of bachelor's theses encompasses two main steps: deconstruction and genre identification. First, the 'big' texts of bachelor's theses were deconstructed into smaller meaningful units basically from an ideational perspectiveexplicit shift in themes (field, in Hallidayan terms), with additional clues from obvious boundary indicators, such as section/chapter headings, and discourse markers (connectors and other meta-textual signals). Second, each of the shorter texts was labelled as an instantiation of a specific elemental genre based on a set of differentiating criteria, including their primary purposes, schematic structures, critical linguistic features, as well as some typological parameters that have been set in previous works by Martin and his colleagues.
To increase the reliability of genre analysis, a guest researcher with extensive experience in SFL genre research was invited to cooperate with the first author of this paper in conducting a pilot analysis. The two coders independently analysed 30% of the corpus -(i.e., 12 bachelor's theses comprising 235 shorter texts) and achieved 92.3% agreement. As for the residual disagreement, either the concerned thesis writer, if available, or a third expert who was familiar with the analytical framework, was consulted for a final solution.
As has been recognised, there is always an on-going list of genres well beyond what functional linguists have already studied or could recognise from their folk rhetoric (Martin & Rose, 2008). Aware of that, we kept an open eye to the theoretical possibility that new genres unexploited before might emerge through the lens of new types of discourse data, and in that case, we would try to label them by ourselves according to their recognisable social purposes and distinctive generic features. From the pilot analysis came out one special text that did not seem to fit any of the initial coding categories. On closer examination, it exhibited recognisable rhetorical patterns with essential similarities with, and some major differences from, the explanation genres, suggestive of a new genre in the same family. It was labelled tentatively by the two coders as analytical explanation. A new code was then added to the operating framework, which need to be consolidated with more similar instances coming in from the full corpus. We shall return to this newly identified genre for further discussion in Section 4.2 below.
The other two types of qualitative data were subject to analysis based on thematic coding, using a constant-comparative method (Glaser, 1978;Strauss & Corbin, 1998). In the first step, all interviews were transcribed and then translated into English. As a form of member-checking, the interview transcripts were referred back to the interviewees to confirm that there was no misinterpretation. Then, we read all the transcripts and documents carefully, noting on any unit or chunks of data with heuristic significance. The recurrent themes and concepts that emerged from this step of open coding were compared back-and-forth to the textual evidences gleaned from the corpus, until the themes were correlated to form a nuanced understanding of the rhetorical values in this thesis writing community. Table 1 presents the overall frequency counts and proportions of all elemental genres and genre families identified in our corpus, which are then graphically displayed in Figure 2 and Figure 3 that follow. In total, the 40 bachelor's theses contained 776 instances of elemental genres, averaging 19.4 cases per thesis. Twenty-two types of elemental genres were identified across the corpus, with no instances being found for observation, recount, news story, personal response, and critical review. To explicate in more detail, canonical examples from the data will be used in cases where we deem it necessary to highlight the manner in which the elemental genre is staged and how it contributes its rhetorical share to achieving the ultimate goal of bachelor's theses. Given the limited space, the texts presented are abridged versions, but the contents remain in the sequence they appear in the original. Key rhetorical eISSN: 2550-2131 ISSN: 1675-8021 311 stages and phases that signal the genres are indicated. Errors or grammatical mistakes in the original copies, if any, are not to be discussed.

. Overall distribution of elemental genres in the corpus of bachelor's theses
It is clear that reports carried the greatest amount of rhetorical weight in the construction of bachelor's theses. Widely used in different parts of theses on a variety of topics, reports comprised nearly half (49.74%) of all the occurrences in the data set, with descriptive reports (32.73%) atop the list followed by classifying reports (11.08%) and compositional reports (5.93%). That reports contained the highest frequency is probably due to that fact that bachelor's thesis writers invested the most rhetorical efforts to characterise the linguistic or cultural phenomena being studied, to inform the readers of the essentials of theories, or to describe the attributes of research instruments; and in cases where more abstract or complicated matters were involved, writers drew up the description by introducing classifications or constituents. On the whole, as Lai and Wang (2018) stated, describing the quality, status and formation of entities is the cornerstone of all forms of academic or scientific activities.
Unsurprisingly, bachelor's thesis writers made heavy use of arguments, making up 26.29% of all cases. Among the three sub-categories, expositions were used most frequently, amounting to 24.48%, second only to descriptive report in the overall ranking, whereas the other two members, i.e., discussions (1.42%) and challenges (0.39%), were of minimal use in the corpus. The most common sites for expositions were the two obligatory parts of a bachelor's thesis -Introduction and Conclusions, where the thesis writers employed expositions either to establish the importance, necessity, or a valued purpose of the research (as shown in Table 2), or to comment on the implications of the research or pinpoint its limitations and weaknesses. By contrast, there were far fewer cases prompting the thesis writer to inspect an issue from opposing perspectives by way of a discussion, or to refute an existing claim via a challenge.
Regardless of the variability in academic levels, 'the development of an argument is regarded as a key feature of successful writing by academics across disciplines' (Wingate, 2012, p. 145). By developing arguments -expositions in particular, the bachelor's thesis writers were expected to demonstrate their powers of independent thinking and to persuade readers to align with their claims. Looking into the institutional context, we found two criterial statements in the Department's assessment rubric that stressed the importance of a sound argument in bachelor's theses, accounting for 30% of the total credits: (1) The thesis has an unambiguous main argument, supported by rigorous reasoning and solid evidences; (2) The thesis conveys original and creative ideas. More studies on the back-translating techniques are in need. Argument Back-translation can better give expression to cultural transmission because the back-translator has to analyze both the original text and the translated text, which is a good way to judge the effect of the cultural transmission and improve the methods to transmit culture through translating. Reinforcement So it is of great necessity to study on the back-translation and give translators some suggestions.
Following arguments, text responses, which are likewise rich in evaluative meanings, made up 8% of the overall corpus; however, they were almost exclusively confined to the 2 literature theses, with only a very few cases in theses of cultural studies and translation, and completely absent from those of applied linguistics. The number of interpretations (5.15%) remarkably exceeded that of reviews (2.84%), perhaps because studies in and of literature were largely engaged in interpretative work, dealing with issues such as themes, characters, or symbolic meanings in a selected literary text, the purpose of which was to rhetorically explain why the selected text was of social, historical, and aesthetic significance. Here, the excerpts in Table 3 represents a typical example of interpretation.

Loss of Self Evaluation
If one is isolated and settles in new surroundings, he will usually do something abnormal. In this novel, almost all kids felt the loss of themselves. Synopsis character recontextualisation event recontextualisation The obvious example was Jack. When Jack went into the forest to find food, he even didn't dare to kill pigs. However, after he overcame his fear, he became addicted in hunting and forgot his responsibility to keep the burning. By killing pigs he could get back to his own values. He succeeded in killing one after another to prove his ability. Jack had lost himself and turned into a totally different person. From then on, the virtuous boy from the civilized world had disappeared; a cruel savage took his place. The tribe killed Simon and Piggy under the leadership of Jack. Jack's loss of himself finally turned him into an actual beast.

Reaffirmation of Evaluation
These kids in Lord of the Flies underwent the loss of themselves in a new society. They lost their own characteristics and did as the others do. The constraint of ethical tradition gradually weakened, and the human nature finally was destroyed.
The analysis showed that reports, arguments and text responses, made up an overwhelming majority (84.02%) of the entire corpus. In contrast, the other four genre families, i.e., explanations, chronicles, procedural genres, and stories, in rank order, played rather auxiliary roles, filling up the remaining 15.98% of the corpus. In what follows, we shall provide details of these auxiliary genres.
Explanations were used occasionally, with 6.44% of the total, to establish causal links between phenomena or events. Factorial explanations and consequential explanations were eISSN: 2550-2131 ISSN: 1675-8021 314 relatively more frequent, accounting for 3.99% and 1.55% respectively. Thesis writers in these cases were driven by a need to unearth the causes leading to a particular cultural or historical event or its observed impacts. In empirical studies, particularly, factorial explanations were used by the writer in order to explain acquired data from surveys or experiments by finding out possible reasons or contributing factors behind. Table 4 is a clear exemplar. To start with the social factor, living in remote rural places which are far from the most developed eastern coastal cities. The only one they can obtain some scientific advice from is the teacher. Factor 2 The next is teachers' individual characters. From Table 2, it can be inferred that the most influential element is their positive personality. Factor 3 The proficient major knowledge and education skills are less important. They have to master some certain knowledge and intelligence level, but once the competence has reached a critical level, it will not have conspicuous effect on their students.
Chronicles were equally sparse, accounting for only 4.64% of the overall corpus, but they played a rather distinct role in the macrogenre by colouring the discourse with shades of 'historical meanings'. Historical recounts were the most common sub-category (3.22%), used to trace how an event of cultural or historical significance happened in the past. By contrast, historical accounts were rarer (0.64%), indicating that thesis writers were not so much concerned with the causal links between historical events as with the sequencing of such events. Meanwhile, biographical recounts, with a relatively low frequency of 0.77%, were seen in texts where the life stories of prominent authors, translators or historical figures were related, as Table 5 illustrates. Commenting on Austen, Edmund Wilson once said: "In around one century and a quarter of the history of English literature, several interesting revolutions occurred to Britain. The renovation of literary taste affected almost all the popularity of writers. But only Shakespeare and Jane Austen are enduring." Adeline Virginia Woolf said: "Among all the great writers, Jane Austen is the most difficult to be caught in the moment of greatness." Procedural genres accounted for 3.87% of the total occurrences of elemental genres, with procedural recounts predominating (2.84%), most notably in theses undertaking empirical studies. The reason for the predominance of procedural recounts is that these empirical studies were already accomplished activities, so the thesis writers were obliged to give, retrospectively, a detailed account of how the research had been done. At the same time, a small number of procedures were also found, mostly at the end of the Introduction, as some thesis writers routinely laid down therein, prospectively, a step-by-step process of how to get the thesis done. Only two instances were found for protocols, in both of which the thesis writers summarised the implications of their research for future translation activities in the form of 'strategies' or 'rules', regulating what to do or not to do for translators.
Finally, stories were the least used, occupying almost negligible proportion (1.03%) in the overall corpus. They were completely exclusive to theses of cultural studies, probably because only writers in this field of study opted for interesting tales from legend or mythology to exemplify their propositional content. It is worth mentioning that the number of exemplum (0.64%) was the highest among the genre family, reflecting the thesis writers' inclination to project or trigger moral judgements related to their topical issues. As we see it, stories, with a primarily entertaining social function, helped to increase the readability of bachelor' theses, and thus created an amusing platform in this stereotypically stodgy site of academic writing. An example of exemplum in Table 6 illustrates this rhetorical effect:

Source and Image of Western Dragon Orientation
It is not unique but has its similar cases. About in the Mid-5th century of Anglo-Saxons, the Britain had a masterpiece which was honored as the national epic of the Anglo-Saxons nation. It was called Beowulf. Incident It's centered on a story in which the hero, Beowulf, struggled to slaughter a fierce monster. Like the creature of Greek tales above, this sort of monster was none other than a dragon that guarded the treasure.

Interpretation
Although Beowulf was badly injured by the dragon's fire, he finally put an end to monster's life through courage and determination.
In summary, the textual spaces in bachelor's theses were furnished by the coconstruction of these elemental genres, which, by means of both number and variety, accounted for the macrogenre' generic complexity, a nuanced understanding of which brought to light the types of social interactions created through the written discourse. Table 7 below summarises the elemental genres at play in bachelor's theses in terms of how (their social functions) and where (their most common sites) they each contribute to the overall complex goals of the macrogenre. Essentially, their use and location in a thesis are not always fixed and invariable. For instance, some genres, like reports, are ubiquitous in theses of all types, while some others, like expositions, though appearing in theses of all types, are more commonly found in 'Introduction' and 'Conclusion' chapters (the only two functional chapters that every thesis must include). Also, there are some genres whose use is rather content-or subject-specific, such as stories in cultural studies and text responses in literary studies, and some others are most narrowly confined to a particular place in a particular type of theses, like procedural recounts in 'Methodology' of empirical studies in applied linguistics. Facing such a variability, what thesis writers need to do is to choose consciously the most appropriate genre at different points in the overall writing processes. Crucially important in making the choice are careful considerations of the nature of their study, the subfield that their selected research topic falls into, and the specific function associated with a particular part or section of the thesis. What we have attempted to offer here, by all means, is eISSN: 2550-2131 ISSN: 1675-8021 316 not a ready-made template that every piece of writing can easily fit into, but rather a pool of rhetorical resources that prospective thesis writers in this discipline can tactically and selectively avail of.
In the section that follows, we focused on a new elemental genre, analytical explanation, emerging from the corpus

'WHAT HAS HAPPENED, IN THIS ASPECT, IS DUE TO THIS REASON': ANALYTICAL EXPLANATIONS AS AN EMERGING ELEMENTAL GENRE
As reported in Section 3.3, a special case emerged in the pilot corpus that attracted the two coders' attention, which resulted in the creation of an additional code, i.e., analytical explanation. The entire corpus yielded 4 instances of this new elemental genre. Though not large in number, these 4 texts were extracted and analysed carefully in order for us to build a refined description. This genre is described and illustrated below. Analytical explanations may begin with an optional Orientation stage that 'locates' or contextualises the phenomenon to be explained and then conclude with an optional Extension stage that functions to draw concluding implications or reflection on the issue. Most importantly, it is the middle obligatory stages that distinguish analytical explanations from the other explanation genres. The Phenomenon stage summarises the outcome to be accounted, although in some cases its role can be performed synoptically by way of section headings or visual auxiliaries (such as tables, figures, or diagrams). In the Explanation stage, one aspect of the outcome is announced and then its cause(s) explained, and this stage is potentially, and necessarily, recursive, construing what Martin & Rose (2008) refer to as an implication sequence. Using the conventional notations of SFL, the schematic structure of analytical explanations is illustrated as below: (Orientation) ^ Phenomenon ^ Explanation: Aspect/Cause ^ (Extension) One example is used to illustrate this genre (Table 8), drawing attention to how particular wordings realised the recursive semantic meanings that characterise the genre and define its stages. In this study, a biophysics corpus is compiled and the basic data statistics and the usage of connector in comparison with that in academic writing, analyzed with both qualitative and quantitative methods. Explanation: Aspect 1 outcome reason On the whole, biophysics corpus has a STTR of 37.91, which is slightly lower than 41.20 in BNC, which is partly because more registers are included in BNC. Explanation: Aspect 2 outcome reason As for POS, the top 10 POS in biophysics corpus are NN, IN, JJ, NNS, DT, MP, CD, CC, VVN and RB. Besides, CC and passive voice are frequently used to increase the objectivity and reliability of the texts. This is because conjunction always represents strong logical connection and passive voice usually takes the object as the subject and put the process of action in the noticeable place, which emphases the truth itself. With 'the' ranking first, most frequently used words belong to functional words. Large number of content words begin to emerge after the top 20 words. A proper explanation of the high frequency of functional words is that their types are very limited, while they have great power to constitute numerous expressions and phrases. Explanation: Aspect 4 outcome reason In terms of connectors, biophysics corpus uses more connectors than academic writing, and the top ten overused connectors in biophysics corpus account for 81% of the total D-value of connectors between two corpora. In semantic category, both corpora have similar tendency in the usage of connectors. This indicates that in comparison with general corpus, biophysics corpus shows the same preference for the category of connectors, while the total usage of connectors in biophysics corpus is far more than that in general corpus.
Noticeably, analytical explanations demonstrate considerable family resemblances with explanation genres, especially factorial explanations. Just as the latter, analytical explanations have to explain how things happen in terms of causes or reasons as their primary social purpose; however, from a typological perspective, the two genres differ significantly in one major aspect: that is, in factorial explanations, the Explanation stage encompasses multiple factors contributing to the Phenomenon in its entirety, whereas in analytical explanation, the Phenomenon is de-composed into its constituting elements or different aspects. This distinction, as shown in Figure 4, is the most crucial criterion for differentiating the two explanation genres, and it influences in turn the way how the two genres unfold rhetorically. Given these refined descriptions, we would propose that the analytical explanation be viewed as an emerging elemental genre, which, to our best knowledge, has remained unaccounted for in the existing scholarship. Presented below is a statement of its rhetorical function: The analytical explanation seeks an all-sided account for a complex problem or phenomenon, by breaking it into its constituting elements or different aspects and then trying to explore the possible reasons behind each of the elements or aspects. It is analytical in the sense that it demands the writer's abilities to see the multi-faceted nature of an issue and thus helps bring people to a horizontally rather than vertically accumulated understanding.
As such, a useful gloss of the analytical explanation is 'What has happened, in this aspect, is due to this reason'.  Findings from genre analysis of 40 bachelor's theses written by Chinese English majors have been presented in this section. Next, we discuss some of the main implications, and moving away from the text into the context, we attempt to tease out the reasons for the choice of genres and the values in this thesis writing community.

BEHIND THE CHOICE OF GENRE: RHETORICAL VALUES IN BACHELOR'S THESIS WRITING COMMUNITY
Drawing on both the textual evidences from the corpus as well as the qualitative data, we identified three core rhetorical values underlying this community of practice. Each will be elucidated in turn in this section.

REPORTS AS THE TRANSMISSION OF EXISTING KNOWLEDGE
In the academic ranks of education, Tardy (2005) argued, students progress gradually from tasks of 'knowledge-telling', in which they write to prove their understanding of existing knowledge, to more complex tasks of 'knowledge-transforming', in which they actively create new knowledge, as advanced academic writers are normally expected to do (p. 325). At the undergraduate level, however, thesis writers are most likely wrestling with issues at the less advanced level, acting as 'knowledge-tellers' rather than 'transformers', having as one of their imperatives the transmission of subject-matter knowledge in appropriate generic forms.
The emphasis on transmitting existing knowledge is evidenced in the Department's Writing Norms of Bachelor's Thesis, which specifies that 'The Introduction should give a comprehensive review of the related studies by previous scholars, presenting the background knowledge of the selected topic.' Likewise, the Assessment Rubric also addresses this issue by integrating the abilities to 'consult existing literature', 'provide sufficient references', and 'analyse all sorts of material' into one domain in the assessment criteria under the label of 'material collection and utilisation', making up 10% of the total credits.
As we have seen from the corpus, reports, whose primary social function is to inform, rather than to entertain or to evaluate, stood as the most appropriate rhetorical vehicle to transmit such received knowledge. Presenting descriptions of the objects being studied, in terms of their defining characteristics, classifications or component parts, most of which were richly documented, reports were in effect a means of 'writing to learn in content areas' (Hirvela, 2011). This argument has found corroboration from our interview data. A thesis writer, whose thesis was on translation of rhetoric in English advertising, confirmed in the interview the role played by reporting genres in transporting the existing knowledge: There are numerous types of rhetorical devices and translation methods, so I classified them and then described one by one in detail. This was how I organised my thoughts in thesis.
…the first two chapters, which mainly included widely-accepted and well-established ideas, such as definitions, linguistic features and rhetorical devices in advertising.' (Writer of Thesis 2018_6) Embracing the existing knowledge also surfaced as a recurring theme in the interviews with thesis advisors. Two of them commented this way (all names used in this paper are pseudonyms): We were much more concerned with the students' ability to search for the relevant literature, and then in the theses, to articulate their understanding clearly. What we would love to see is their understanding of the literature. (Wendy, a thesis advisor) eISSN: 2550-2131 ISSN: 1675-8021 320 When we supervise the students, one of our chief concerns is to help them select and sort out reference materials related to their selected topics, and ideally, express in their own words how they understand the materials. (Sophie, a thesis advisor) Another advisor informant, perhaps due to her relatively increased familiarity with the concept of genre, pinpointed the close relationship between the genres of reports and the transmission of existing knowledge: For bachelor's theses, I think, reports and arguments are perhaps more important, relatively, than the other genres.
Because the primary goal of bachelor's theses is for students to transfer what they have learnt from a wide range of reference materials, reading, analysing and integrating them into a research product. Based on that, students also have to try to draw their own conclusions. For the dual purposes, reports and arguments, I reckon, may have more weight to carry in their writing. (Grace, a thesis advisor) It is interesting to note that, alongside reports, this interview excerpt also pointed to the significance of arguments.

ARGUMENTS/TEXT RESPONSES AS THE PROJECTION OF AUTHORIAL SELF
Authorial self is an important concept in the theoretical discussions of academic writer identity (Clark & Ivanič, 1997;Ivanič, 1998). According to Clark & Ivanič (1997, p.152), the authorial self involves 'the textual evidence of writers' feeling of authoritativeness and sense of themselves as authors.' Integrating this concept of self with SFL -the appraisal theory (Martin & White, 2005;White, 2015) in particular, McKinley (2018) described the authorial self as an indication of the writer's sense of an exigency to write, recognised at any point when writers put forward opinion or stance in establishing their own argument.
In our corpus, the frequent use of arguments and text responses can be understood as a means for the writers of bachelor's theses to project such an authorial self. Both performing a social function on the evaluative end -to evaluate either an idea or a text (Rose, 2010), arguments and text responses thus accumulate meaning potential that individual writers can avail of for personal interpretation, analysis, and evaluations, as well as for endorsement or critiques of others' opinions or works.
This ability to argue critically is highly valued in the thesis writing community. One of the thesis advisors interviewed regarded the ability to evaluate opinions, which he referred to as 'critical thinking', as the primary function of bachelor's theses: The writing of bachelor's theses can help them develop a critical thinking ability, to look at problems from different angles. This ability also keeps them from accepting others' opinions too easily. To enhance the ability to think critically, I must say, is the primary function of bachelor's theses. (Brown, a thesis advisor) In the later talk-around-text session, this thesis advisor went on to assert that the rhetorical structure of a bachelor's thesis was an argument at its most macro level -a macro-argument in SFL terms, capturing the nature of writing a bachelor's thesis as a persuasive endeavour.
At the most macro level, a bachelor's thesis is in essence an argument, but it comprises various parts. You see, in a bachelor's thesis, only two parts, Introduction and Conclusion, are necessarily arguments. The rest can vary. Seen from a global perspective, it is definitely an argument. (Brown, a thesis advisor) The importance of being able to argue critically was also recognised by some thesis writers, as two of them commented in the interviews: …but more important is to form our own ideas and perspectives from the vast sea of existing literature. (Writer of Thesis 2018_2) Actually, I think, to argue and then to explain are the key themes in a thesis. (Writer of Thesis 2018_7) One of the thesis advisors emphasised, additionally, the arguments in bachelor's theses had to be well grounded in theories in the discipline, or based on careful implementation of research procedures and solid empirical evidences. In other words, thesis writers are expected to build an argument using evidences from discipline-appropriate sources rather than from their personal and emotional standpoints.
However, life in the community was not without 'unsuccessful attempts to argue', which may lead to a loss of the writer's voice and authorial self, as one of the thesis writers reflected in the interviews: I should have included arguments of my own in Chapter 3, but I hardly did. I don't think I was really able to… (Writer of Thesis 2018_6) Of course, the construct of academic writer's voice or authorial self is a complex, elusive issue and is obviously well beyond the scope of a single section in the current paper. What we see through the lens of this construct is the work that argument genres were actually doing, or expected to do, in the discourse of bachelor's theses. Arguments, and text responses in theses of literary studies, can therefore be deemed to be an appropriate choice of genres where a display of the authorial self is called for. This is the second rhetorical value in this local community of practice.

ANCILLARY GENRES AS THE POSITIONING OF A CONSTELLATION OF SOCIAL ROLES
Beyond the reporting and evaluating genres, an array of genres was found simultaneously at play to lesser but varying degrees, giving rise to a multitude of social roles that 'versatile' thesis writers were prompted to perform in the macrogenre of bachelor's theses.
Specifically, the hybridisation of these ancillary genres opened up an ample discursive space in between those informative and evaluative meanings, so that the thesis writers could be positioned often simultaneously as storytellers (by means of stories) or historians (by chronicles), or positioned to seek resolutions or explanations (by explanations), or positioned to get things done by observing step-by-step procedures or strict rules (by procedural genres). As such, the community of practice is moulded by the way in which the writers adopt these stances, and how they accept, decline or tailor these positionings, to eventually become a rhetorically flexible member of the writing community.

CONCLUSION AND IMPLICATIONS
Our genre analysis of 40 bachelor's theses offered a rich account of the generic complexity involved in this academic macrogenre. It revealed, specifically, how various elemental genres were jointly constructed, and then how a new genre, analytical explanation, emerged from this macrogenre embedded in undergraduate research. Furthermore, drawing on multiple sources of data, we looked beyond the written texts into the wider context, which allowed us to tease out the rhetorical values held by core participants in the thesis writing community. Discussions along these two lines, as have presented in this paper, led us to a broad conclusion that, if the writing of bachelor's theses were to be characterised as a value-laden 'human activity' (Russell, 2001), the configuration of genres did provide a discursive means to materialise these values.
The main findings from our study lend useful guidance to thesis writers and thesis advisors in similar EFL contexts. A better understanding of the generic composition of bachelor's theses will facilitate future final-year students to approach this high-stakes academic macrogenre with a sharpened 'rhetorical vision'. This rhetorical vision empowers students not only with an increased knowledge about how varieties of elemental genres can be navigated and flexibly appropriated to reach the diverse, complex goals in bachelor's theses, in compliance with the rhetorical values in the local writing community, but also a heightened meta-awareness of genre throughout the whole processes of writing. Likewise, such a genre knowledge and genre awareness can hopefully equip thesis advisors with a 'pedagogical metalanguage' (Rose, 2017) which they can use when they sit down with their advisees and give instruction or feedback to their writing.
The present study has several implications for future research. First, our study has only examined bachelor's theses which were highly assessed; future studies can benefit by including and comparing theses across the spectrum, in order to reveal the correlation between the deployment of genres and the judgement of the merits of the theses. Second, with a corpus in which the sub-fields of the discipline (i.e., literature, culture, translation or applied linguistics) are better represented, for example, via stratified random sampling, future research can further investigate the nuanced variations across the sub-fields, a move which was beyond the scope of this study but would provide crucial information about how the relevant research topics would exert influence on the disciplinary writers' choice of genres. Third, as undergraduate education is a preparation for more advanced learning and research, comparing bachelor's theses with higher-level genres, such as master's or doctoral theses or published research articles written by expert writers in close disciplines would be a valuable line of inquiry. Finally, comparisons can also be made to equivalent macrogenres in other disciplines, which points to future applications of the approach we have developed here. Such studies can enhance our knowledge about the varied rhetorical values in disciplinary communities and bring into light how the deployment of genres shapes and is shaped by the discipline-specific cultures.