Found in Translation: Late Bilinguals Do Automatically Activate Their Native Language When They Are Not Using It

Area of Science:

• Psycholinguistics
• Cognitive Neuroscience
• Computational Linguistics

Background:

• Late bilinguals sometimes show native language (L1) effects when processing their second language (L2).
• Costa et al. (2017) proposed L1-independent priming in bilinguals, suggesting L2 word links formed via L1 form overlap explain these effects.

Purpose of the Study:

• To investigate the validity of Costa et al.'s model for explaining L1-independent priming in late bilinguals.
• To re-examine the role of L1 activation in L2 processing using computational modeling.

Main Methods:

• Building upon Costa et al.'s simulation code for a Hebbian learning model.
• Modifying the model's inter-language connection severing to test L1 influence on L2 activations.

Main Results:

• The Costa et al. model only reproduced L1-independent priming under specific L1 disconnection conditions.
• Bidirectional severing of inter-language connections eliminated the influence of L1 form overlap on L2 activations within the model.

Conclusions:

• The model's ability to explain L1-independent priming is highly dependent on implementation details.
• Further theoretical and modeling efforts are needed to fully account for language non-selective activation in late bilinguals.