محمدمهدی شرفی

This study investigated the relationship between Iranian high school students’ Universal Grammar knowledge and their performance on such standardized general English proficiency tests as PET and FCE internationally administered by Cambridge University. To this end, 108 students were randomly chosen from some high schools located in Malayer from Hamedan. Since this study was correlational in nature, and descriptive and hypothesis-testing by definition, the research participants were given no treatment. Three tests were administered to them instead. To measure UG knowledge, a researcher-made UG test was given to all participants. This test which was made both reliable and valid included pied-piping and preposition stranding principle, binding principle, pro-drop parameter, that-trace effect, projection principle, resumptive pronoun and subjacency principle. To assess the participants’ general English proficiency, PET and FCE tests were run. All participants took the three tests consecutively at two-week intervals after they were given clear and detailed instructions. The findings were manifold. Firstly, there was a significant relationship between UG knowledge and performance on the proficiency tests. Secondly, there existed no significant difference between the proficiency tests as far as the UG test was concerned. Finally, the UG test scores were, through the Cubic regression model, proved to predict the scores gained on both proficiency tests. Most importantly, this study led to some suggestions regarding the learner variables and the under-explored issue of integration of generative SLA and language testing, more specifically standardized general English proficiency tests.