Children are sensitive to statistical regularities in speech and likely use these regularities when learning their native language. A central goal of current research is to understand which statistical regularities support different aspects of language acquisition and processing. In the current work we explore phonological and semantic similarity effects on early lexical acquisition. Using a computational model, behavioral findings from word learning studies are simulated and explored. With this model we demonstrate that acquisition can be facilitated by the distinctiveness of individual lexical mappings.