A fundamental weakness shared by second language syllabi is that they have been based on their authors' assumptions about language learning and have lacked an empirically supported, psycholinguistic grounding. The following article will review two major traditions in syllabus design which share this weakness. Underlying one tradition is the assumption that second language structures which are the most different from the learner's L1 are also the most difficult to learn, and therefore should be given strongest emphasis in the syllabus. In the other tradition, it is assumed that there is a direct relationship between linguistic complexity and learning difficulty, and that the syllabus, therefore, should present target structures to the learner in an order of increasing linguistic complexity. This article will re-examine the assumptions underlying these two traditions in syllabus design in light of recent findings from second language acquisition research.