ABSTRACT
Despite growing interests in the acquisition of Chinese orthography, few studies have modeled the acquisition process using connectionist networks. This study uses a self-organizing connectionist model to simulate children’s learning of Chinese characters. There are two major goals of our study: (1) To evaluate the degree to which connectionist models can inform us of the complex structural and processing properties of the Chinese orthography. One of the most difficult tasks in achieving this goal is how to faithfully capture the orthographic similarities of Chinese characters. We derived our character representations on the basis of analyzing a large-scale character database that can be readily mapped to school children’s orthographic acquisition. (2) To test the utility of self-organizing neural networks in orthographic acquisition. Most previous connectionist models of orthographic processing have relied on the use of feed-forward networks. Results from our simulations present positive evidence for both of our goals. In particular, we show that our model demonstrates early regularity effects and frequency effects in the acquisition of Chinese characters, matching up with acquisition patterns from empirical research.
