Examining Structural Constraints in Bilingual Language Processing and Speech Production

Open Access
Author:
Navarro-torres, Christian A
Graduate Program:
Psychology
Degree:
Master of Science
Document Type:
Master Thesis
Date of Defense:
January 15, 2016
Committee Members:
  • Judith Fran Kroll, Thesis Advisor
  • Paola Eulalia Dussias, Thesis Advisor
Keywords:
  • bilingualism
  • cognition
  • language production
  • language use
  • language processing
Abstract:
One of the most important findings in studies of bilingual language processing is that bilinguals co-activate information from both languages across a variety of contexts and domains. While previous studies have reported syntactic co-activation/interaction, very few have examined co-activation of lexical-syntactic information in production and its relation to domain-general cognition. Here we investigate whether bilingualism modulates the use of subtle syntactic constraints in the L2 via a-adjectives (e.g., afraid, alive), which have been shown to resist prenominal use despite their adjectival status (Boyd & Goldberg, 2011). English-monolinguals, Spanish-English bilinguals (immersed in their L2), and German-English bilinguals (immersed in their L1) performed a sentence production task that elicited the use of prenominal and relative clause constructions. The task yielded three main findings. First, while all groups exhibited the a-adjective constraint, bilinguals were overall more stable in their preference. A follow-up analysis revealed that this pattern was related to errors on filler trials: those who prenominalized 3rd person singular verbs (the sings* boy…) were more likely to also prenominalize a-adjectives, and this effect was found to be strongest in a subgroup of monolinguals. Crucially, bilinguals as a whole were less prone to generating these errors. Given the task’s level of difficulty, we argue that bilinguals may have performed the task differentially using a proactive strategy. Second, Spanish-English bilinguals were slower overall producing correct sentences, potentially reflecting inhibition of L1 syntactic information, but German-English bilinguals showed similar latencies to monolinguals, possibly indicating less cross-language interference. Finally, sentences containing a-adjectives were overall produced slower, even when compared to filler trials involving relative clauses, suggesting activation that both lexical and syntactic information are activated at the onset of planning.