Distributional Learning for Other High-Level Structures?
Given our findings regarding the relationship between distributional information and category learning, it seems reasonable to explore the role of distributional learning for other types of high-level language structures. One next step is to take our results on grammatical category learning and extend them to the acquisition hierarchical phrase structure. Some work has explored the usefulness of distributional information for phrase structure acquisition (Takahashi & Lidz, 2008; Thompson & Newport, 2007), but the interplay of syntactic category acquisition and hierarchical phrase structure acquisition is relatively unexplored. Artificial grammar learning experiments provide an excellent method for examining the types of information that are crucial to the successful acquisition of these higher-level language structures.