You are viewing the SNL 2018 Archive Website. For the latest information, see the Current Website.

Poster A68, Thursday, August 16, 10:15 am – 12:00 pm, Room 2000AB

Bilingual Recruitment of Inhibitory Control While Translating

Jamie Renna1, Yazmin Medina1, Ksenija Marinkovic1,2, Katherine J. Midgley1, Phillip J. Holcomb1;1San Diego State University, 2University of California, San Diego

Bilingualism requires the management of two languages within one linguistic system. Cognates, or words with similar orthography and semantic representation across two or more languages, are valuable for studying how bilinguals navigate communication while controlling multiple languages. For instance, Christoffels et al. (2007) observed a cognate facilitation effect during a picture-naming task, indicating a non-target language was activated and facilitated processing during production of words in the target language. The present study further explored the question of non-target language activation by utilizing false cognates, which are words similar in form across languages but different in meaning. If both languages are activated at the presentation of a word, perhaps the different semantic representations of false cognates are also concurrently activated and rather than facilitation the processing of false cognates would instead result in competition. We hypothesized that evidence for competition would come in the form of an interference effect whereby false cognates would be more difficult to produce than non-cognates and real cognates. To test this, we collected electroencephalography (EEG) data while English/Spanish bilinguals completed a translation-production task. From this EEG data, we extracted event-related potentials (ERPs). Participants were instructed to verbally translate words presented visually on a computer monitor and ERPs were time-locked to the presentation of each word. English and Spanish words were randomly intermixed and included 100 non-cognates (e.g., English “dog” and Spanish “pero”), 100 false cognates (e.g., English “sunrise”, which translates to “amancer” in Spanish, and Spanish “sonrisa” which translates to “smile” in English), and 100 cognates (e.g., English “list” and Spanish “lista”) with 50 of each type in each language. We predicted that translating false cognates would result in increased effort due to the need to inhibit the alternative meaning of the word in the target language. We expect this would result in a larger ERP negativity for false cognates compared to non-cognates and real cognates. Consistent with this prediction, in frontal electrode sites there was a small increase in an N400-like negativity for false cognates compared to cognates in the 400–600 ms time window. While this pattern is consistent with the results of Christoffels et al.’s (2007) cognate facilitation findings, we found an even larger divergence among the word types between 600 and 800 ms, with false cognates showing greater negativity than both non-cognates and cognates. Together these results are consistent with our hypothesis that when multiple meanings are activated, increased cognitive demand is required by speakers to suppress the unwanted meaning in the target language.

Topic Area: Multilingualism