Morphosyntax is Easy , Discourse / Pragmatics is Hard

Turkish is a highly inflected, agglutinative language in which subject NPs are never overtly casemarked, but object NPs often are. Although all six possible word orders are grammatical when objects are overtly casemarked, only SOV and OVS word orders are grammatical when objects lack overt casemarking. In imitation experiments, Batman-Ratyosyan and Stromswold (1999) found that 4and 5-year-old Turkish children performed equally well on uninflected SOV and OVS sentences, whereas 3-year olds performed significantly worse on OVS sentences. However, in a comprehension experiment, Stromswold and Batman-Ratyosyan (1999) found that only 3-year olds did significantly better than chance on uninflected OVS sentences. One possible explanation for these finding is that the older children knew enough about Turkish discourse and pragmatics that they found uninflected OVS sentences (in which objects are indefinite/plural, focused and topicalized, see section 1.2) somewhat infelicitous without appropriate discourse context. In this study we investigate how Turkish children use word order, casemarking and discourse-context to determine the thematic roles of sentential constituents. We found that children did best on SOV sentences, and that casemarking only helped with OVS sentences. Although older children did slightly better when context was given, 2-year-old children did significantly worse when context was given. This could reflect 2-year olds’ incomplete linguistic knowledge or inadequate exposure to language-specific rules of discourse/pragmatics. Alternatively, it could be the result of 2-year olds not understanding other people’s intentions and knowledge states.