zum Inhalt springen

Morphosyntactic prediction in early bilingual children

Figen Karaca1, Susanne Brouwer1, Sharon Unsworth1, & Falk Huettig1,2

1 Radboud University, Centre for Language Studies, 2 Max Planck Institute for Psycholinguistics

Previous research has mainly focused on prediction skills of late bilinguals, while prediction skills of early bilinguals who acquired their two languages simultaneously have hardly been examined (Karaca et al., 2021; Schlenter, 2022). The present study therefore investigated prediction skills of early bilingual children based on case-marking cues in comparison to monolingual children. In a visual-world eye-tracking paradigm, 49 Turkish-speaking monolingual children (Mage=7.1, SD=0.5) and 30 Turkish-Dutch bilingual children (Mage=7.6, SD=1.1) listened to Turkish sentences in which case-marking on the first NP (e.g., rabbitNOMINATIVE/ACCUSATIVE) and the verb (e.g., eat) position (sentence-medial, sentence-final) were manipulated, while looking at images of the first NP, an agent and a patient (e.g., rabbit, fox, carrot) (c.f. Özge et al., 2019).

Analyses of looks to the agent image showed a significant interaction between Time (in the predictive window), Condition (accusative, nominative), and Group (monolingual, bilingual) in verb-final (β=-0.18, SE=0.03, z=-6.50, p<.001) and verb-medial sentences (β=-0.23, SE=0.02, z=-9.54, p<.001). When the prediction effect was examined separately in each group, a significant interaction between Time and Condition was found for both groups in verb-final and verb-medial sentences (see Figure 1). These results suggest that, under the influence of a non-case-marking-language (i.e., Dutch), early bilingual children, though to a different extent than monolingual children, are able to generate predictions based on morphosyntactic cues that are transparent and acquired early.

Figure 1: Agent preference of monolingual and bilingual children in verb-final and verb-medial sentences

Note: Agent preference in 50 ms time bins averaged across participants and across trials. The error bars indicate the standard error of the mean across participants. The red line is for the accusative condition and the blue line is for the nominative condition. Positive values on the y-axis indicate preference for the agent image and negative values indicate preference for the patient image, while 0 indicates no preference for either image. The red rectangles mark the predictive time windows.

  • Karaca, F., Brouwer, S., Unsworth, S., & Huettig, F. (2021) Prediction in bilingual children: The missing piece of the puzzle. In E. Kaan & T. Grüter (Eds.), Prediction in second language processing and learning (pp. 116–137). Amsterdam, Philadelphia: John Benjamins
  • Özge, D., Küntay, A., & Snedeker, J. (2019). Why wait for the verb? Turkish speaking children use case markers for incremental language comprehension. Cognition, 183, 152–180. https://doi.org/10.1016/j.cognition.2018.10.026
  • Schlenter, J. (2022). Prediction in bilingual sentence processing: How prediction differs in a later learned language from a first language. Bilingualism: Language and Cognition, 1–15. https://doi.org/10.1017/S1366728922000736