This project investigates second language (L2) learners' processing of four types of Chinese relative clauses crossing extraction types and demonstrative-classifier (DCl) positions. Using a word order judgment task with a whole-sentence reading technique, the study also discusses how psycholinguistic theories bear explanatory power in L2 data. An overall preference for DCl-first structures and an advantage of DCl-subject relative clauses over the other three structures were found. Results were largely compatible with the filler-gap domain theory and indicated a weak subject-gap advantage. These motivations are subject to influences from other factors, and a multi-constraint proposal was proposed.