Semantic and translation priming from a first language to a second and back: Making sense of the findings

The present study investigated cross-language priming effects with unique noncognate translation pairs. Unbalanced Dutch (first language [L1])-English (second language [L2]) bilinguals performed a lexical decision task in a masked priming paradigm. The results of two experiments showed significant translation priming from L I to L2 (meisje-GIRL) and from L2 to L I (girl-MEISJE), using two different stimulus onset asynchronies (SOAs) (250 and 100 msec). Although translation priming from L I to L2 was significantly stronger than priming from L2 to L 1, the latter was significant as well. Two further experiments with the same word targets showed significant cross-language semantic priming in both directions (jongen [boy]-GIRL; boy-MEISJE [GIRL]) and for both SOAs. These data suggest that L1 and L2 are represented by means of a similar lexico-semantic architecture in which L2 words are also able to rapidly activate semantic information, although to a lesser extent than L I words are able to. This is consistent with models assuming quantitative rather than qualitative differences between L I and L2 representations