Within the field of L2 metaphoric competence (MC) research, Low’s (1988) and Littlemore and Low’s (2006a, 2006b) metaphor-related skills and (sub)competences have existed for 29 and 11 years respectively, but have never been elicited or used to develop tests. Consequently, the extent to which they are underpinned by more fundamental (sub)constructs is unclear. With a few exceptions (e.g., Littlemore, 2001), L2 MC tests to date have been limited in scope (e.g., Aleshtar & Dowlatabadi, 2014; Azuma, 2005; Hashemian & Nezhad, 2007; Zhao, Yu, & Yang, 2014). Available research shows that L2 MC correlates with L2 vocabulary knowledge and proficiency (Aleshtar & Dowlatabadi, 2014; Zhao et al., 2014), but negligibly with time spent in an L2 immersion setting (Azuma, 2005). However, the ability of these measures to predict L2 MC is unknown, as is the change in the receptive/productive correlation strength as L2 proficiency increases.
In response to these gaps, a large battery of L2 MC tests aimed at eliciting Low’s (1988) and Littlemore and Low’s (2006a, 2006b) constructs was developed and administered to 112 NNSs of English (L1 Chinese) and 31 English NSs, along with vocabulary knowledge and (NNSs only) general proficiency tests. Data cleaning showed inherent, operationalisation problems. Exploratory Factor Analysis revealed four metaphor-related factors, with MANOVA and independent samples t-tests showing statistical NNS and NS differences for only one of these: English Grammatical Metaphoric Competence. Multiple regression revealed that the Oxford Online Placement Test best predicted L2 receptive MC, whereas L2 vocabulary depth measured by the Word Associates Test (Read, 1998) best predicted L2 productive MC. Time spent living in the UK had no predictive power, and the receptive/productive correlation weakened with increased L2 proficiency. Implications for theory, test development, the transferability of models and predictors (e.g., to NNSs with other L1s) and EFL teaching are discussed.
