A corpus-assisted exploration of grammatical metaphor across populations: successful student writers, expert authors, and EFL learners

Grammatical metaphor (GM) represents arguably the key linguistic resource for the construction of meaning in written academic discourse (Schleppegrell, 2004). However, despite its importance, there are a number of ways in which GM remains poorly understood. For example, much of our knowledge of GM use extends from small-scale studies of a limited range of populations. Our knowledge of how GM is best taught is particularly restricted as it relates to EFL and EAP contexts. GM is not well represented in pedagogical research or in the prevailing approaches to EAP instruction. Although there is a broad consensus regarding the value of explicit instruction of GM in EAP contexts (e.g., Fang et al., 2006; G. Hu & Perez, 2022), we lack a clear idea of what form this instruction should take.

This thesis adopts a combination of dissertation-by-publication and monograph approaches as it seeks to address these gaps in our knowledge. It does this through a corpus-assisted analysis of GM use across three populations: successful student writers, published authors, and EFL learners. The analysis of the writing of the first two of these populations is conducted at a larger scale than has previously characterised GM research. GM use by successful student writers is examined through a 164,925-word sample of the British Academic Written English Corpus (BAWE). The analysis of GM use by published authors deploys a 200,000-word corpus derived from published research articles. The findings from these two stages of the study reflect both consistency and variation. Broad patterns of GM use were found to be largely consistent, both across disciplines and populations, while considerable variation was found in the specific GMs deployed. The findings of these first two stages informed the development of a number of analytical tools that it is hoped will prove invaluable in future GM-related research and pedagogy. One is a core list of GMs that occur with some regularity across disciplines and both populations. More significantly, the data extracted from the BAWE and the research article corpus were used to develop a Grammatical Metaphor Word List (GML), a 4,574-item list of the GMs identified in the two corpora.

The findings of the first two stages were also instrumental in the development of the third stage of this study. This stage adopts a reduced scale to examine the effects of explicit instruction on GM use by a group of EFL learners at a Japanese university. Broadly speaking, this stage of the study finds a strong, largely positive, effect for instruction on GM use. During this investigation, texts written by a group of six students are more closely examined to draw specific inferences about the effects of instruction, and to make recommendations about what form successful GM-focused pedagogy might take.

It is hoped that the findings of this study will not only improve our understanding of GM use across populations and disciplines, but that through the GML in particular, this thesis will facilitate more effective and more efficient future research and instruction of grammatical metaphor.