On the Properties of Syntactic Priming: A L2 Hierarchical Bayesian Model Approach

A range of empirical factors have been identified in the literature as interacting with the strength of syntactic priming: the lexical boost, the inverse frequency effect, and the asymmetrical decay. This study explores how these factors can be represented within a general learning framework called the hierarchical Bayesian model (HBM), utilizing data from the K-English Textbook corpus. The HBM conceptualizes syntactic knowledge as a hierarchical structure of syntactic statistics, which is continually updated through Bayesian inference based on the language experience (Xu and Futrell 2024). Given this background, the current research aims to investigate the underlying mechanism of syntactic priming from a different angle using statistical learning. After building the L2 HBM, two simulations are conducted employing Pickering and Branigan’s (1998) English ditransitive materials. In so doing, we demonstrate that the L2 HBM successfully captures the aforementioned properties of syntactic priming, as a previous study reported. To account for these observed factors simultaneously, we support the claim that empirical properties of syntactic priming are realized in the cognitive model architecture.