Abstract
Three experiments were carried out to investigate whether the bilingual dual coding theory is fit for early- and late-balanced bilinguals and second language learners. All experiments required them to encode a mixed list of their first and second language (L1 and L2) words by translating L1 words and copying L2 words with either imagery or non-imagery instruction. The subjects were then asked unexpectedly to free recall L2 words they had generated. The results of Experiment 1 and 2 showed that item recall was best for the imagery condition, intermediate for the translation with non-imagery condition, and worst for the copying with non-imagery condition in an approximately 3:2:1 ratio. The additive effect reflected in the above ratio was interpreted as supporting the bilingual dual coding theory which assumes the independent and interconnected storage systems for bilingual words memory. In Experiment 3, however, the additive effect was only partially observed. It was suggested that the bilingual dual coding theory was not applicable to second language learners who had not developed sufficient interconnections between the second language system and the imagery system.
