Vocabulary Development in English and Chinese: A Comparative Study with Self-Organizing Neural Networks

In this paper we present a self-organizing neural network model to simulate the early vocabulary development in English and Chinese. We focus on how the different lexical composition patterns in the two languages can emerge, develop and change when the learner acquires an increasing number of words. Our results suggest that certain lexical characteristics in the linguistic input (e.g., word frequency and length) play significant roles in the presence or absence of given patterns. Our study presents a dynamic developmental picture for early lexical acquisition, which is dependent on the joint contributions of mechanisms of learning and characteristics of the learning environment.