| In four experiments we explored the degree to which a bilingual's two languages are active prior to language production utilizing a variant of the Peterson and Savoy (1998) production paradigm. In Experiments 1 and 2, the primary production task is picture naming. In Experiments 3 and 4, the primary production task is word translation. In each experiment, the primary task is interrupted on 33% of the trials with a single word probe, which had to be named as quickly as possible. Word probes were either identical to the name of the picture, its translation equivalent, a word that is phonologically similar to the translation equivalent, a semantically related word, or an unrelated control. In Experiments 1 and 3, production occurs in L1 (Dutch). In Experiments 2 and 4, production occurs in L2 (English). In addition to exploring the relative activity of both languages we also compared picture-naming trials to word translation trials. Previous research has suggested that these two tasks are conceptually mediated however we hypothesize that these tasks may affect language selection differently.; Consistent with previous findings we observed cross language identity/translation effects in both picture naming and translation trials suggesting that both languages are active prior to language production. However, a comparison of the picture naming and translation results revealed that these two tasks affect language selection quite differently. In translation language selection occurs earlier than in picture naming because the word stimulus provides an inherent cue to what language to produce in, which is consistent with the language cue hypothesis. |
