An Exploratory Study on Authorial (In)Visibility across Postgraduate Academic Writing: Dilemma of Developing a Personal and/or Impersonal Authorial Self

The writers of any scientific community are inherently expected to fulfil some agreed-upon discourse conventions of the academic discourse community (Molino, 2010) in the sense of creating a successful dialogic interaction through their texts. In line with this, Akbas (2014b) raised the question of "how and to what to extent writers foreground their explicit manifestations or hide their personal projections with impersonal forms" (p. 56). Considering the fact that academic writing is closely linked to the representation of authorial self (Hyland, 2002) and the voice of the postgraduates has received relatively less attention, in this paper, we explored the notion of explicit (via I and we-based instances) and implicit (via passive and impersonal instances) representation of postgraduates as the novice writers in the Social Sciences; namely, Turkish native speakers, Turkish speakers of English and English native speakers. Therefore, the focus of the paper shall be on the vari ations of personal (first person pronouns) or impersonal (agentless passives and inanimate subjects) uses of authorial references as well as their discourse functions in the postgraduate writing. In total, 90 successfully-completed dissertations of three postgraduate groups were randomly selected to compile the corpus of the study, and a corpus-informed discourse analysis approach was applied in the identification of choices of authorial representation in this genre. Following an extensive manual analysis of the texts from the corpus for each group, a list of explicit and implicit authorial references was extracted from sample texts to be explored in the analysis of the whole corpus. As was applied by Fløttum (2012), during the identification of authorial references, all verbs collocating with the explicit or implicit authorial references were examined carefully to see if the references performed author visibility in the texts. The quantitative analysis clearly showed that Turkish L1 and Turkish writers of English preferred to build mainly an impersonal impression over what they were present ing to the reader by employing a greater number of implicit authorial references whereas English L1 writers chose to create a more self-prominent academic prose. The qualitative analysis provided some evidence to argue that three groups employed explicit or implicit authorial references to accomplish particular discourse acts (i.e. guiding readers through the texts, elaborating an argument & making