Exemplification in research articles: Structural, semantic and metadiscursive properties across disciplines.

Examples are forms of code glosses similar to reformulations that help readers understand writers’ intended meanings (Hyland, 2007) and contribute to the process of argumentation in texts (Triki, 2017). This paper investigates examples in research articles across the so-called soft and hard disciplines and aims to structurally and semantically explore the link between the building units of examples and how they could constrain the choices made across disciplines. To reach this end, the study investigates examples in a corpus of 80 research articles that cover four disciplines in the ‘soft’ sciences and four disciplines in the ‘hard’ sciences. Annotation was based on an automatic search of potential candidates then a manual annotation was performed using the UAM Corpus Tool (O’Donnell, 2008). Results suggest that (1) the structural features of exemplified units impose constraints on the choice of markers and exemplifying units (2) exemplifying clauses are more elaborative than exemplifying nominal groups and (3) the equivalence between the units exemplified and those exemplifying seems to be governed by the degree to which exemplifying units are expanding or compacting. The main conclusion is that the soft-hard distinction between academic disciplines is not always the major parameter triggering writers’ choices.