Cross-Language Priming Asymmetries in Lexical Decision and Episodic Recognition

Masked translation priming between languages with different scripts exhibits a marked asymmetry in lexical decision, with much stronger priming from L1 to L2 than from L2 to L1. This finding was confirmed in a lexical decision task with Chinese-English bilinguals who were late learners of English. Following a suggestion made by Bradley (1991), the experiment was repeated using a speeded episodic recognition task. Participants studied Chinese words, and then were tested in an old/new classification task in which Chinese target words were primed by masked English translation equivalents. Significant priming was obtained for old items, not for new items. However, no priming was obtained when lexical decision was used. Unexpectedly, the episodic task showed a reverse asymmetry, since L1-L2 priming was not obtained with this task, although strong effects were obtained for lexical decision. A possible explanation for this pattern of results is that knowledge of L2 lexical items is represented episodically for late learners.