Grammatical Complexity of Grammatical Metaphor: A Dependency Relation-Based Approach

The extant quantitative research within the domain of grammatical metaphors (GMs) in Systemic Functional Linguistics (SFL) has predominantly relied on the normalized frequency of these structures across various registers to gauge the complexity of textual materials. This frequency-based approach, however, falls short in capturing the intricacies inherent to individual GMs, thereby rendering it inadequate for a comparative analysis of the lexical or grammatical/syntactic complexity among different GMs or even the same GM within divergent linguistic contexts. The present study posits that the grammatical complexity associated with GMs may be more accurately quantified by examining the dependency distance, diversity, and strength of these structures in relation to other syntactically interconnected elements within the same linguistic context. The findings are encapsulated as follows. Firstly, the utility of dependency relations as a metric for assessing the grammatical complexity of GMs is substantiated. Secondly, the deployment of dependency relations in the prediction of EFL writing quality reveals that while dependency diversity does not serve as a robust predictor of EFL writing quality, both dependency distance and strength of GMs emerge as efficacious predictors of EFL writing quality. This research will shed light on future inquiries into the theoretical underpinnings of GMs, the integration of dependency relations with additional metrics for the prediction of EFL writing quality, and the refinement of manual scoring protocols, or even the development of automated scoring mechanisms.