The nature of cross-language lexical activation in sentence context: A pscyholinguistic analysis

Open Access
Schwartz, Ana Isabel
Graduate Program:
Doctor of Philosophy
Document Type:
Date of Defense:
September 24, 2003
Committee Members:
  • Judith Fran Kroll, Committee Chair
  • Paola Eulalia Dussias, Committee Member
  • Sandra J. Savignon, Committee Member
  • Cathleen M Moore, Committee Member
  • Rick Owen Gilmore, Committee Member
  • Chip Gerfen, Committee Member
  • bilingualism
  • lexical access
  • sentence processing
The goal of the present study was to examine the cognitive nature of second language (L2) lexical processing in sentence context. Recent psycholinguistic research has provided evidence that bilingual lexical access is language non-selective in nature and that when bilinguals recognize words, information from both languages is activated in parallel. These non-selective findings have shaped current models of bilingual lexical activation and our understanding of how the bilingual lexicon is organized. However, these findings have been primarily based on isolated word recognition. Since bilinguals function in context-rich environments, it is important to understand to what degree information provided by context modulates language non-selectivity. In the present study we examined the influence of sentence context by comparing bilinguals’ lexical processes in out-of-context, and sentence-context tasks. Our goal was to determine whether the top-down process of L2 sentence comprehension interact directly with the bottom- up processes of lexical access and whether this interaction would constrain language non-selectivity. In the first set of experiments we examined bilinguals’ L2 word recognition performance for language-ambiguous words (e.g., cognates and interlingual homographs) in two out-of-context tasks (word naming and lexical decision). These experiments replicated the effects of cross- language activation that have been observed in the literature. In the second set of experiments the same critical words were inserted in high constraint and low constraint sentences. The findings demonstrated that effects of language non-selective access persisted in low constraint sentences. This suggests that the language context provided by a sentence is not sufficient to constrain non-selectivity. However, in high constraint sentences the effects of cross-language activation were eliminated for bilinguals with relatively high L2 comprehension skills. The results suggest that sentence context can indeed constrain non-selectivity when there is sufficient semantic information and processing resources to efficiently suppress the non-target native language. Implications for current models of bilingual lexical access are discussed.