Bulut, T. 1, 2, 3 , Uysal, H. 4 & Wu, D. 2, 3
1 Özyegin University, School of Languages
2 National Central University, Institute of Cognitive Neuroscience
3 National Yang-Ming University, Laboratories for Cognitive Neuroscience
4 University of Florida, Department of Linguistics
Previous research shows that across many languages such as English, object relative clauses (ORC) are acquired later than subject relatives (SRC). Frequency, structural and linear distance between dependents are among the factors put forward to explain this asymmetry. However, it is not certain if and how typologically different languages exhibit such a contrast. We used Turkish, which is a head-final language with canonical SOV word-order and overt case marking, to investigate this issue. Since the frequency of SRC is greater in both child speech and child-directed speech as found in a Turkish corpus study, it is predicted that SRC should be easier to acquire and process compared to ORC. On the other hand, accounts based on linear distance between dependents predict ORC advantage, since this distance is shorter in ORC. With two experimental groups (preschoolers and primary-schoolers) and a control group (adults), the present study aimed to shed light on this debate by means of a sentence-picture matching task. The accuracy results show that in line with findings from English, children performed better at SRC than ORC. The results are compatible with frequency accounts and highlight the role that statistical learning plays in acquisition of linguistic structures.