Language dominance, educational context, and the trilingual lexicon: a large-scale masked translation priming study

The present study employed a masked translation prime paradigm to investigate the nature of the trilingual mental lexicon. It was found that L3 (English) was primed only by L2 (Mandarin) and not by L1 (Korean), which further analysis showed to not be the result of L2 being dominant over L1. It was suggested that a formally learned L3 can be associated with only one of L1 or L2, and this association is influenced by the language in which L3 instruction was conducted. The further finding that L2 primed L1 but not vice versa, indicated that the former was more strongly represented than the latter and, again, it was apparent that the standard measure of dominance could not explain this. Two explanations were entertained, one centreing on the possibility that it is the language of formal education in general that determines the relative strength of the language representation, and the other on the possibility that dominance is modality-specific.