This study investigates lexical development in second language (L2) learning using a complex network method within the framework of complex dynamic system theory (CDST). Analyzing written output from L2 Chinese learners of varying proficiency and language backgrounds, the study differentiates proficiency levels using a bi-gram lexical network model. Lexical network indices (network density, clusters) provide a deeper understanding of L2 proficiency than traditional indices (average word length, hapax legomena percentage, though Guiraud index proves robust). The study reveals consistent complex network characteristics across proficiency levels and highlights the value of word association features beyond word frequency in characterizing interlanguage systems. The authors argue for integrating both lexical frequency and network features in CDST studies of L2 lexical development.

Heng Chen