The problem of segmenting continuous speech is a great challenge for native and nonnative (L2) listeners alike. Speech signal, unlike the presence of gaps in written text, often do not contain breaks at word edges. Even when breaks occur in the form of pauses, they do not coincide with perceived word boundaries. Listeners can solve the segmentation problem by utilizing a variety of cues including lexical, acoustic, and prosodic cues. Language typology may influence the integration and weighted importance of the cues. The current dissertation will examine English L2 segmentation by learners with Korean, Mandarin, or Spanish L1 backgrounds. In each of the five proposed experiments, a pair of cues will be compared: 1) stress cues versus lexical knowledge; 2) stress cues versus acoustic cues; 3) phonotactic cues versus lexical knowledge; 4) phonotactic cues versus semantic cues; and 5) stress cues versus sentence context. The goal of the dissertation is to propose a hierarchical framework that describes the differential weights given to segmentation cues by L2 learners of English. This dissertation may have significant implications for pedagogical practices as findings will inform learners and teachers to selectively direct their attention to cues in the L2 that are not utilized by the learners in L1 segmentation.