Do Formal Stance Strategies Reveal Disciplinary Variation in Professional Scientific Writing?

Stance in academic discourse has been extensively studied, with numerous investigations indicating that its expression varies across disciplines, depending on the authors’ intention to either enhance or diminish their voice or presence (e.g., It seems fairly certain vs. This is based on the belief that…). This paper hypothesises that stance can be viewed as a strictly structural or formal linguistic mechanism in academic discourse, which can optimally determine disciplinary variation. The novelty of this study lies in the hypothesis that formal features of stance alone can identify academic disciplines, without relying on the meaning conveyed by the features. To demonstrate this, this paper focuses on the linguistic expression of stance in hard- and soft-science articles. The corpus of soft and hard scientific writing consists of research articles published in leading peer-reviewed journals in eight disciplines (chemistry, physics, engineering, mathematics, business studies, history, linguistics and political science) and comprises approximately 1.6 million words. The assessment of the realisation of stance in the aforementioned scientific disciplines is carried out by quantifying a range of grammatical (e.g., modal verb groups and embedded complement clauses) and lexical (boosters, hedges, and self-mention expressions) features suggested in the literature. The frequencies of the features are statistically modelled by means of, firstly, a multivariate regression analysis that determines the set of features whose contribution to the hard- versus soft-science variation is significant and, secondly, a clustering technique that groups similar disciplines based on exclusively the frequencies of the significant stance features. Clustering very successfully reveals a neat classification of the eight disciplines under investigation into two major clusters corresponding to the initial categorisation of the writings into the hard- and soft-science types. This suggests that the meaning conveyed by the stance features is dispensable for the purpose of disciplinary categorisation.