Transfer of constraints at the initial state of phonological acquisition

Abstract

This thesis investigates language transfer in bilinguals at the beginning of the learning of a novel language. Participants are speakers with L1 English and L2 French, or L1 French and L2 English. The task consists of having participants produce coda clusters that are allowed in Russian. Results show that both groups performed well on clusters that are allowed in English and French, but poorly on those that are not found in either language. Optimality Theory analyses show that the English L1 group only transfers their English knowledge, while the French L1 group transfers both English and French knowledge. This is shown through the repair strategies used, as the English L1 group tended to repair using medial epenthesis, and the French L1 group used a combination of medial and word-final epenthesis. These results show support for the Linguistic Proximity Model, as it accounts for the combination of sources for transfer. To explain these patterns of transfer, it is hypothesized that English is transferred because of psychotypology, or possibly because English is the most restrictive language as it has the smallest amount of allowable coda clusters. Under the Subset Principle, the most restrictive language is optimal for facilitating learning of novel languages.