1. Introduction

This article introduces the concept of psychological vocabulary and its network structure in the brains of second language learners, pointing out that traditional research mainly explores its internal structure through word association types and is influenced by second language proficiency. The research of network science has revealed the characteristics of second language word networks, such as density and modular attributes, and found that the structure of second language word networks is influenced by factors such as individual acquisition experiences. Although there have been studies, there has been no attention paid to the structure of English second language word networks among high-level Chinese learners and their influence on second language proficiency, and the existing research is not detailed enough in terms of quantitative methods. This study aims to use network science technology and manual classification methods to systematically investigate the structure of Chinese learners' English second language word networks from three levels: macro, meso, and micro. It explores whether it is a complex network, compares it with native language word networks, and the impact of second language proficiency on word network structure.

2. Research methods

This paper introduces the methods for studying the network structure of second language psychological vocabulary of Chinese learners in English. The research subjects include high - and low-level English second language groups and English native language groups, consisting of 32 English major students, 40 high school students, and 31 American college students, respectively. The research tool is an English word association test questionnaire, which includes 49 stimulating words, of which 40 are from middle school English textbooks and 9 are low familiarity words. The subjects completed the vocabulary association test within 10 minutes. The study constructed empirical and random networks, processed data using Excel and Matlab software, extracted network parameters using the BCT toolbox, and visualized the network using Gephi software. The random network was constructed using the Erd ö s-R é nyi model and 1000 simulations were conducted to obtain the parameter mean.

3. Results and Discussion

Through paired sample t-test and SPSS 26.0 software analysis, a comparison was made between the Chinese second language psychological vocabulary network and the random network. It was found that the empirical network had significantly lower average shortest path and average clustering coefficient than the random network, indicating that the English second language vocabulary network has complex network characteristics. The effect size analysis shows that the difference in average clustering coefficient is more significant. The study also found that the word networks constructed by second language learners at different levels have complex network characteristics, and second language proficiency does not affect the essential properties of the word network.

4. Conclusion

This chapter summarizes the main findings of the research on the network structure of Chinese learners' second language psychological vocabulary: the second language word network has complex network characteristics, and there is no significant difference from the native language word network at the macro level. The modularity level at the meso level is weak, and there are qualitative differences in the micro level connection patterns; Second language proficiency affects microstructure. Constructed an optimized second language psychological vocabulary complex network model, pointed out the limitations of the research and future research directions.