The Influence of Cultural Context on Language Activation in Korean-English Bilinguals

Both languages for bilinguals are jointly activated even when performance is clearly restricted to one.The present study investigated the role of cultural cues on the relative level of joint linguistic activation.Twenty-two Korean-English bilinguals were presented with a picture and an audio cue and indicated via button press whether the heard label named the depicted object while EEG was recorded.In the critical blocks, the pictures represented exemplars that were more typically English or Korean, even though both exemplars take the same name in both languages (e.g., North American soup vs. Korean soup).English or Korean labels for the same set of pictures were presented in separate blocks.Reaction times were significantly faster for trials in which the auditory stimulus correctly named the object and the language matched the cultural bias.Providing the correct label in either language significantly attenuated the N400.A late positive component (LPC) was present for trials in which the label was correct, and was more positive when viewing Korean exemplars with English audio.No differences were seen when either English or Korean pictures were paired with Korean auditory stimuli.Therefore, effects of cultural context and semantic integration appear to be separate. Nonselectivity of SpeechThe initial evidence for nonselectivity of linguistic access came from Poulisse (1997) who showed that in Dutch-English bilinguals, Dutch (their first language, L1) interfered with English production (L2), producing speech errors.These errors were attributed to a functional frequency effect in which the infrequently used L2 was likely to be influenced by the stronger L1.While this pointed to a nonselectivity of speech in that the L1 was involved in a context requiring L2 selection, the data did not specifically determine at what stage in speech planning the L1 had its effect on the L2.To determine the point in speech planning when this influence occurs, Hermans, Bongaerts, DeBot, and Schreuder (1998) utilised a 'cross-language picture-word Stroop paradigm'.Dutch-English bilinguals with Dutch as L1 were asked to name pictures in English, with distractor words being presented either before, after, or during picture presentation.In one study these distractor words were presented in English and in another study they were presented in Dutch.Distractors were either semantically or phonologically related to the picture, phonologically similar to the Dutch translation (Phono-Dutch), or unrelated.Consistent with monolingual research in which semantically related words interfere with production and phonologically similar words facilitate production (Schriefers, Meyer, and Levelt, 1990;Kuipers, La Heij, and Costa, 2006), they found interference when the English semantic distractor (e.g.'valley') was presented close in time to the picture ('mountain'), and facilitation from the English phonological distractor ('mouth') at a greater range of presentation times.The most critical result showed interference with naming response times when the distractor was phonologically similar to the Dutch