The impact of first and second language exposure on learning second language constructions

Abstract

We study how the learning of argument structure constructions in a second language (L2) is affected by two basic input properties often discussed in literature – the amount of input and the time of L2 onset. To isolate the impact of the two factors on learning, we use a computational model that simulates bilingual construction learning. In the first two experiments we manipulate the sheer amount of L2 exposure, both in absolute and in relative terms (that is, in relation to the amount of L1 exposure). The results show that higher cumulative amount of L2 exposure leads to higher performance. In the third experiment we manipulate the prior amount of L1 input before the L2 onset (that is, the time of L2 onset). Given equal exposure, we find no negative effect of the later onset on learners’ performance. This has implications for theories of order of acquisition and bilingual construction learning.

Modelling verb selection within argument structure constructions

“…The model of construction learning demonstrated similarity to human data in terms of U-shaped learning patterns, use of syntactic bootstrapping (in both production and comprehension), phenomena of over-generalisation and recovery (Alishahi & Stevenson, 2008. Finally, the model was adapted for simulating bilingual construction learning, demonstrating effects of amount of input similar to those in human learning (Matusevych et al, 2015b).…”

“…One objective of the current study is to simulate EOR's experiments using the computational model of argument structure construction learning (Alishahi & Stevenson, 2008;Matusevych, Alishahi, & Backus, 2015b). The second objective is to test whether the findings of EOR still hold after addressing the two methodological issues described above; the computational model is particularly helpful in this respect.…”

“…We use the data sets available from Matusevych et al (2015b); here we briefly outline how they were obtained.…”

Comparing Computational Cognitive Models of Generalization in a Language Acquisition Task

“…In addition, both models use batch processing, failing to address how learning to generalize across an alternation might be achieved incrementally. Alishahi and Stevenson (2008) presents an incremental Bayesian model shown to capture various aspects of verb argument structure acquisition (Alishahi and Pyykkonen, 2011;Barak et al, 2012Barak et al, , 2013bMatusevych et al, 2016), but the model is unable to mimic alternation learning behaviour. Barak et al (2014) extends this constructionlearning model to incrementally learn both constructions and classes of alternating verbs, and show the role of the classes in learning the dative.…”

Taxing the bilingual mind: Effects of simultaneous interpreting experience on verbal and executive mechanisms

“…Far from being static, such systems can exhibit significant enhancements depending on how early and competently they are put to use. Increased L2 exposure has been linked to better performance in tasks tapping linguistic (Matusevych, Alishahi & Backus, 2015) and non-linguistic (Bosma, Hoekstra, Versloot & Blom, 2017) skills, alongside plastic changes in critical brain regions (Grant, Fang & Li, 2015). Also, evidence from lexical (García, 2015a) and executive (Linck, Osthus, Koeth & Bunting, 2014) tasks shows faster and/or more accurate performance in high- than low-proficiency bilinguals.…”