Bilingual word learning: Recognizing novel words in context

Abstract

Background

This study examined the effects of semantic and language variability on novel word learning in bilinguals. We investigated whether varied semantic contexts improve novel word learning in both the dominant and non-dominant language, particularly for novel words across a range of cross-language orthographic overlap.

Methods

Two within-subjects experiments were conducted with English-Spanish (n = 48) and Brazilian Portuguese-English bilinguals (n = 68). Participants studied rare words in repeated or varied semantic contexts across different language conditions (dominant-only, non-dominant-only, or both languages). Reading times for sentences during the study session, accuracy and response times in a semantic relatedness task, measured word learning.

Results

Sentence reading times decreased with each presentation, especially in repeated contexts and when words were presented in the dominant language. Cross-language orthographic overlap significantly moderated reading times of the whole sentence when novel target words were studied in both languages (translation equivalents). The post-test semantic relatedness task presented more complex findings. English-Spanish bilinguals benefited more from words studied in varied semantic contexts, while Brazilian Portuguese-English bilinguals showed an advantage for repeated contexts, likely due to differing task completion strategies and language use patterns between the groups.

Conclusions

While some findings aligned with the Context Variability Hypothesis, other findings revealed more nuanced effects of semantic and language variability. Findings were interpreted within an instance-based theoretical framework of word learning. Language variability across study encounters may provide additional retrieval cues, facilitating novel word learning in bilinguals. However, the effects are moderated by task demands and specific language-pairs.