This empirical study examines the learning problem of the argument structure of psych predicates such as «The dog frightens John» and the related V-ing adjectives such as «The dog is frightening to John». The problem is theoretically interesting because of the marked nature of the thematic role mapping of these sentences in relation to the principle of the Uniformity of Theta Assignment Hypothesis (UTAH). The problem is highly relevant to our understanding of second language acquisition, as this is known to be a prevalent difficulty among language learners. The author has framed the learning problem within a coherent parametric framework drawing on a sophisticated critical review of the syntax/semantics literature and theories of L2 development. The author has specifically developed a theory, the «Semantic Salience Hierarchy Model» (SSHM), to explain the learning process. The significance of the model is not confined only to this particular study, as the issues related to the L2 acquisition of other causative verbs can also be examined within this model. The findings of this study also bear implications to TESOL.