Quantifying cross-linguistic influence with a computational model

Abstract

Cross-linguistic influence (CLI) is one of the key phenomena in bilingual and second language learning. We propose a method for quantifying CLI in the use of linguistic constructions with the help of a computational model, which acquires constructions in two languages from bilingual input. We focus on the acquisition of case-marking cues in Russian and German and simulate two experiments that employ a picture-choice task tapping into the mechanisms of sentence interpretation. Our model yields behavioral patterns similar to human, and these patterns can be explained by the amount of CLI: the negative CLI in high amounts leads to the misinterpretation of participant roles in Russian and German object-verb-subject sentences. Finally, we make two novel predictions about the acquisition of case-marking cues in Russian and German. Most importantly, our simulations suggest that the high degree of positive CLI may facilitate the interpretation of object-verb-subject sentences.