Shuyi Amelia Sun

Recent corpus‐based studies suggest generative AI systems such as ChatGPT can approximate human patterns of interpersonal metadiscourse in argumentative writing. These conclusions are typically based on AI‐generated texts produced using a single prompt, despite AI output being highly sensitive to even minor changes in prompt specification. This study examines prompt specificity as an elicitation variable in the production of metadiscourse. Using Hyland's interpersonal metadiscourse framework, we first compared university students’ argumentative essays from the LOCNESS corpus with GPT‐generated essays produced in response to a general essay (baseline) prompt specifying role, genre, topic, length and persuasive purpose, but providing no explicit guidance regarding metadiscourse use. We then generated a second GPT corpus using a metadiscourse‐inducing prompt that explicitly requested interactive, interactional, and engagement features that were found to be underrepresented in the initial GPT corpus relative to the LOCNESS corpus. The baseline prompt produced lower frequencies of transitions, hedges, self‐mention, attitude markers and engagement resources than human writing, while the metadiscourse‐inducing prompt substantially increased several target features over the baseline and human conditions. Some categories converged with baseline frequencies, but even where convergence occurred, these were often realised in more formulaic ways. The findings show prompt specificity significantly shapes both the distribution and rhetorical realisation of metadiscourse in AI‐generated writing. More broadly, the results highlight the importance of treating prompt specification as a key methodological variable when interpreting linguistic patterns in AI‐generated texts.

The ability to achieve social interaction is both a key feature of research writing and an important aspect of advanced academic literacy. It can be seen in how doctoral students employ rhetorical resources to acknowledge limitations in thesis writing while securing a positive view of the research. Negation is one of the crucial interactional options, but less explored in English for Academic Purposes (EAP) literature. In this study, we drew on the appraisal theory to see negation as a disclaim marker that engaged with alternative positions and employed corpus analysis to examine the forms and functions of negation in the ‘limitations’ section of doctoral theses. To better understand how student writers exploit negation to achieve the rhetorical end, we further explored co-articulations of negation with other appraisal resources. A corpus-based analysis of 100 doctoral theses by Chinese and American doctoral students in applied linguistics showed that American students made significantly more use of negation, especially pairing negation with engagement and graduation resources. We attribute the difference to genre and culture norms and also raise pedagogical implications on the cultivation of students’ rhetorical awareness and the classroom teaching of research writing.

As an innovative and systematic genre in the academic community, Ph.D. theses have been heatedly researched in the field of English for Academic Purposes. Although research on the functional and formal features of Ph.D. theses has been abundant, their stylometric traits regarding textual activity have not been explored. Accordingly, this study explored the textual activity of Ph.D. theses and its dynamic changes across natural sciences, social sciences and humanities. A total of 150 Ph.D. theses (50 from each discipline) were analyzed, and the Q and χ^2 values were calculated to determine the textual activity of theses as well as its dynamic changes with the progression of texts. The results showed that, although the theses were found to be active in general, significant differences across disciplines do exist, in that the theses in natural sciences and humanities were more active while those in social sciences were more likely to lean towards the descriptive mode. This study has implications for widening the scope of cross-disciplinary academic genre analyses from an innovative quantitative linguistic perspective.

The past few decades have witnessed an upsurge of scholarly interest in the generic descriptions of PhD theses following Swales’ seminalGenre Analysis. Fitted within the Create A Research Space (CARS) model, the thesis introduction plays a key role in justifying research originality/significance, where novice writers engage with academic communities through “establishing a research territory” (Move 1), “establishing a niche” (Move 2), and “occupying the niche” (Move 3). As the hinge of the CARS model, Move 2 (hereinafter EN) is of strategic importance as it enables writers to “sell” their ideas by pointing to the gap/niche in the “marketplace” of previous research, which is typically realized through the co-occurrences ofnegationalongside other interpersonal language resources.Negation, as adisclaimmarker within Martin and White’sappraisalframework, is a prominent linguistic indicator of EN. Nevertheless, little research has systematically examined the use ofnegationin ENs of PhD thesis introductions. Accordingly, the study investigatednegationvia theappraisalframework addressing subtypes ofnegation(disalignment,cautious detachment,unfulfilled expectation) within ENs in the introduction sections of 120 PhD theses. The results showed thatdisalignmentis the most frequent subtype ofnegation, while “not” and “no” are commonly used as indicators ofnegation. Our findings also revealed intriguing co-occurrences ofnegationsub-categories alongside other relevantappraisalresources. The corpus-informed results are expected to shed light on the nature and practice of PhD theses that educators may take into account during thesis writing instruction.

This study employed a text mining method to investigate the lexical features and their dynamic changes of PhD theses across the natural sciences, social sciences and humanities. Four quantitative indices, i.e. TTR, h-point, R1 and writer’s view, were employed to analyze 150 PhD theses (50 theses from each discipline). Although h-point and writer’s view were found counter-intuitively to show insignificant variation across disciplines, the results of TTR and R1 did reveal sharp contrasts between theses in humanities and natural sciences. While the second half of humanities theses showed a significantly higher level of lexical diversity, indicated by higher TTR, theses in natural sciences tended to be richer in content words in the first half, indicated by a higher R1. Meanwhile, theses in social sciences seemed to be more moderate, with features lying in the middle position. This study has implications not only for the widening of applications of quantitative linguistic methods but also for academic writing (especially PhD thesis writing) instruction and practice.