Eye Movement Measures of Cross-Language Activation During Reading in Bilingual Children and Adults: A Focus on Neighborhood Density Effects

Thematic Section: Modulators of cross-language influences in learning and processing

cross-language influence, transfer, immersion, morpho-syntax, lexicon

Veronica Whitfor, University of New Brunswick & University of Western Ontario
Marc F. Joanisse, University of New Brunswick & University of Western Ontario

Several eye-tracking studies have examined cross-language activation during reading in bilinguals. This work has found that bilinguals have an integrated memory system, wherein both their languages are concurrently represented and non-selectively accessed. Although this work has advanced our understanding of bilingual language processing, it has almost exclusively focused on cognate and/or interlingual homograph processing (which involves overlapping representations across languages) among young adult university students (which may lack generalizability).

Here, we employed a more conservative measure: cross-language orthographic neighborhood density to examine cross-language activation in bilingual children and adults across their known languages. Thirty-three English-French bilingual children (aged 7 to 12) and 30 English-French bilingual adults (aged 18 to 21) read four ~100-word texts in their L1 and L2 while they were eye-tracked (Whitford & Joanisse, 2018). Each word was coded for total cross-language orthographic neighborhood density (an updated measure that includes substitution, addition, and deletion neighbors), length, frequency, and predictability.

We had three main findings. First, cross-language neighborhood density effects were facilitatory across the L1 and L2. Words were easier to process when they had many vs. few cross-language orthographic neighbors. Second, cross-language neighborhood density effects were more pronounced in the L2 vs. L1. In the L2, they were observed during both early- and late-stage reading; whereas in the L1, they were only observed during late-stage reading. Third, the magnitude of cross-language neighborhood density effects was larger in children vs. adults across the L1 and L2 (and both reading stages). Children found words with few cross-language orthographic neighbors especially difficult to process.

Our findings suggest that cross-language neighbourhood density facilitates reading behavior, especially in the L2 and among children. Thus, the activation of orthographically similar word forms (i.e., neighbours) from bilinguals’ more proficient language (L1) eases both early and late stages of L2 word processing, particularly among developing readers.