Self-mention, the way authors present themselves in their writing, is a crucial interactional resource in academic discourse. It shapes authorial identity and engagement with readers. Previous research has categorized self-mention into three sub-categories: first-person pronouns (FPP), third-person nouns (TPN), and abstract subjects (AS). While existing research has explored synchronic aspects and some diachronic features of self-mention, a comprehensive diachronic and cross-disciplinary study is lacking. This study aims to address this gap by examining changes in self-mention and its sub-categories in soft science (Applied Linguistics and Sociology) and hard science (Biology and Electrical Engineering) articles from 1970 to 2019. The research questions focus on the diachronic changes in the use of self-mention and its sub-categories and the cross-disciplinary differences in their use over the 50-year period.

A substantial body of research has investigated self-mention in academic writing. Studies have linked self-mention to various discourse functions, including explaining procedures, stating findings, justifying propositions, and building connections with readers. Cross-linguistic studies have revealed higher frequencies of first-person pronouns in English compared to other languages. Cross-cultural research has shown differences in self-mention usage between L1 and L2 English writers. Previous work has also explored cross-disciplinary variations, with soft sciences generally exhibiting higher frequencies of self-mention than hard sciences, although the specific forms and locations of self-mention vary across disciplines. However, existing diachronic studies have primarily focused on first-person pronouns, neglecting the broader picture encompassing TPN and AS. This study addresses this limitation by analyzing all three categories across various disciplines over a long time frame.

The corpus consisted of 1200 single-authored research articles from 12 leading journals in four disciplines: Applied Linguistics and Sociology (soft science) and Biology and Electrical Engineering (hard science). Three journals per discipline were selected based on impact factors. Articles were selected from 1970 to 2019, with two articles per journal per year. The 50-year period was divided into five 10-year periods. Only the Introduction, Method, Result, Discussion, and Conclusion sections were analyzed. Liu's (2011) classification framework was adopted for categorizing self-mention, with additions based on a preliminary analysis. AntConc 3.4.3 was used for frequency counts, and manual checks ensured accuracy. Normalized frequencies per 10,000 words were calculated to account for article length variations. Mixed effects models in R 4.3.0 were used to analyze the data, treating writers as random effects and discipline and year as fixed effects. Both normalized frequencies and percentages of each sub-category within self-mention were analyzed to investigate diachronic trends within and across disciplines.

The analysis revealed significant diachronic and cross-disciplinary differences in self-mention. Overall, self-mention showed an increasing linear trend from 1970 to 2019. However, the sub-categories showed non-linear changes: first-person pronouns and abstract subjects increased, while third-person nouns decreased. In soft sciences, self-mention decreased significantly over time, with a notable decline from the 2000s to the 2010s. This decrease was mainly driven by a reduction in first-person pronouns. In contrast, hard sciences showed a significant increase in self-mention, particularly first-person pronouns, reaching a peak in the 2010s. The percentage of first-person pronouns decreased in soft sciences (from 88.8% to 77.8%) and increased in hard sciences (from 85.9% to 91.0%). Third-person nouns remained relatively stable in soft sciences but decreased in hard sciences. Abstract subjects increased in both soft and hard sciences, but their percentage increased dramatically in soft sciences and decreased in hard sciences. These findings suggest that soft science writers are adopting more subtle self-presentation strategies, while hard science writers are becoming more direct in their self-representation.

The findings challenge the traditional view of soft science as more personal and subjective than hard science. The decreasing trend of self-mention in soft science might reflect an increased awareness of the need to appear objective and avoid potential biases. The increasing trend in hard science could reflect a shift towards greater authorial visibility and a recognition of the importance of emphasizing individual contributions. The changing use of first-person pronouns across disciplines reflects different approaches to establishing authority and engaging with readers. Soft science writers appear to favor more indirect forms of self-presentation, while hard science writers favor more direct forms. The varying trends in the use of abstract subjects likely reflect discipline-specific norms and preferences for balancing authorial presence and objectivity.

This study provides a diachronic and cross-disciplinary overview of self-mention in English research articles over 50 years. The contrasting trends in self-mention usage between soft and hard sciences highlight the evolving norms and conventions of academic writing. Future research could explore the discourse functions of self-mention in greater depth and investigate the influence of social and cognitive factors on authorial choices. Qualitative approaches, such as author interviews, could provide valuable insights into the motivations and perspectives behind these observed trends.

The study's limitations include the selection of journals based on impact factor, which might not fully represent all journals within each discipline. The focus on single-authored articles might limit the generalizability of the findings to multi-authored works. Future studies could consider a broader range of journals and investigate the impact of authorship on self-mention patterns.