A GENRE ANALYSIS OF LITERATURE REVIEWS IN DOCTORAL THESES A Genre Analysis of Literature Reviews in Doctoral Theses 對於博士論文中文獻綜述之體裁分析 Submitted to Department of English and Communication 英文與傳播學系 in Partial Fulfillment of the Requirements for the Degree of

Abstract Research into thesis-writing began about a decade ago. However, theorizing and research into the task of literature reviewing (LR) to date is still underdeveloped. The two-part genre analysis presented in this thesis is an attempt to fill the theoretical and research voids. The first part of the study is a textual analysis (also called the thin analysis), which investigates the schematic pattern of LR chapters in doctoral theses organized in the traditional format (i.e., Introduction -Literature reviewMethodology -Results -Discussion; abbreviated as ILrMRD). Based on the assumption that LRs and Introductions in research writing share similar rhetorical purposes-as alluded to in thesis instructional literature, the analysis employed, as a starting analytical framework, the 3-move CARS model (Swales 1990) and especially that posited for thesis introductions (Bunton 2001). The LR chapters analyzed are drawn from 20 theses which were produced by a group of Chinese doctoral students based in Hong Kong and cover a wide range of social science and humanity topics. It was found that many of the LR chapters display an Introduction-Body-Conclusion structure. The body segments of most chapters comprise thematic sections which display highly cyclical move patterns that suggest the presence of the 3 moves and their respective elements postulated in the CARS model. The occurrence of Move 1 and Move 2 is notably higher than that of Move 3. There is also a preponderance of Move 1-2 pairing. Elements within each of the moves do not co-occur regularly. When they do co-occur, they do not appear in any predictable order. Three new elements (making confirmative claims, relevancy claiming and abstracting-synthesizing theoretical frameworks /theoretical positions) were also identified in some instances of Move 2. Taken together, the findings suggest that the CARS scheme for introductions may not be entirely applicable to describe the patterns found in the LR chapters. A revised model is thus postulated in the present thesis. The second part of the study is a thick analysis of literature reviewing, which is a response to the recent calls for extending genre inquiries to probe into various processes of text production (Bhatia 1993(Bhatia , 2004 Brant 1990; Devitt 1993). It examines how doctoral students select literature for reviewing (e.g., themes, specific authors, specific sources and specific theoretical frameworks), which constitutes a major step of constructing LRs. One of the aims of the thick analysis is study how technical events such as the writing of LR and research activities implicate (construct) the selection of readings. This aim of analysis is motivated by calls from research writing theorists to deconstruct the demarcation view about reading, researching and writing and to restore the nexus among the three activities in theorizing about research writing. Part of the analysis is cognitive in its orientation and acknowledges doctoral students' agency in the selection of literature for reviewing. The study is complemented by an examination of the social processes involved in reviewing. The exploration has its theoretical underpinning derived from social constructivist theory and in particular Lave and Wenger's (1991) situated learning theory. It looks into how students develop their cognition of 'core literature' for reading by taking into consideration the social milieu in which doctoral students carry out their research studies. ii Stories of RS and RLR were collected from sixteen doctoral student informants. The stories reveal that many of the informants developed their awareness of the key literature while working on the other parts of their studies, which include the pilot studies, data collection, data analysis and drafting of the literature review chapters for major official documents such as the proposal, the qualifying report and the thesis. Literature reviewing apparently served different purposes at various stages of their study journeys. Most of the students began the task as reading to learn for their studies (RS) through which they gained preliminary understanding about the research topics regarding their research focuses, conceptual contours, methodological concerns characteristic of the areas of study, and research design. Reading at this stage was the least focused, which began to resolve at the commencement of various research activities such as pilot studies, data collection and data analysis, in which some of the students grew aware of their immaturely formulated hypotheses, inadequacies in the adopted theoretical frameworks, or underdeveloped operational constructs. Realization of the shortcomings made the students become more critical about their own research and the literature they had reviewed, which at the same time steered the students away from the core literature consulted earlier and moved them to embark on new themes for reviewing. Based on the recounted experiences provided by the student informants, the thesis argues that reading during the initial and research stages of study is less directed at the writing of LR. As the stories reveal, reading for the literature review chapter (RLR) became most marked and most intensive when the students started writing their LRs (WLR) usually during a time when institution-imposed deadlines of submission of major qualifying documents approached. Their LR drafts provided 'heuristic' forms that helped them identify gaps of information which guided reading for specific details. RLR during the final stage of the study journeys, as commented by some of the completing or graduated students, also served the purpose of updating citations in the draft literature review. The stories also show that knowledge of what to read was partly acquired from experienced members in the fields of the students' studies through guided participation in various parts of the students' research studies. These members were primarily the students' supervisors and occasionally panel members, who oversaw the students' progress. Many students reported obtaining reading lists from their supervisors. Some recounted receiving instruction while discussing their research progress or written outputs with their supervisors or panel members at various stages of their studies. The stories suggest the importance of the students' interim outputs as springboards for RS and RLR supervision. Several informants provided accounts of acquiring theoretical frameworks, crucial readings and key authors' names for their RS from extrinsic networks of experts (Kaufer & Geisler 1989) whom the students met at conferences and seminars, during their participation in the supervisor's Project or during the experts' scholarly visits to their departments. The accounts suggest that peripheral participation in academic activities and establishing contacts with members of the field of study plays a crucial role in facilitating students' RS. As stories iii provided by two of the informants also reveal, resorting to extrinsic networks can sometimes be implicated by the mismatch between the supervisor's expertise and the subject matter of the student's study. The study offers rich insights into the complexity involved in literature reviewing as a rhetorical textual product and a socio-cognitive process, from which pedagogical and research implications can be drawn