Error-based learning and lexical competition in word production: Evidence from multilingual naming

Area of Science:

• Psycholinguistics
• Cognitive Science
• Second Language Acquisition

Background:

• Lexical selection is the process of retrieving the correct word during speech production.
• Understanding the mechanisms of learning and competition in lexical selection is crucial for explaining language processing.
• Previous research has explored error-based learning and competitive selection, but their interplay requires further investigation.

Purpose of the Study:

• To investigate whether learning associated with lexical selection is primarily error-based.
• To determine if lexical selection operates through a competitive mechanism.
• To assess the after-effects of word production on subsequent semantic competitor retrieval, considering varying degrees of error.

Main Methods:

• Participants engaged in a two-part naming task involving pictures and words in different languages.
• Part A involved naming stimuli in one language, followed by Part B where participants named the same (old set) and new sets of stimuli in another language.
• Reaction times (RTs) for naming were measured to assess retrieval costs for previously produced items (translation equivalents) versus novel items.

Main Results:

• Retrieval costs for previously named items (translation equivalents) were significantly larger than for new items.
• This retrieval cost was greater when participants initially named words in a higher-error language (L2) compared to a lower-error language (L1), supporting error-based learning.
• When naming occurred in a high-error language (L3), the subsequent cost for naming in another high-error language (L2 vs. L1) was amplified, indicating competitive lexical selection.

Conclusions:

• The findings support an error-based learning model for lexical selection, where errors encountered during word production influence subsequent retrieval.
• Evidence suggests that lexical selection is a competitive process, particularly evident in multilingual contexts with differential error rates.
• These results contribute to a deeper understanding of how the brain learns and selects words, especially in second and subsequent language acquisition.