Although most words have more than one meaning, the mechanisms underlying new-meaning learning have been understudied. This one-to-many mapping poses even greater challenges for second language learners. The present study examined the behavioral mechanisms underlying new-meaning learning among non-native speakers by focusing on the effects of word familiarity, an approximate measure of lexical quality. We found that learning new meanings for more familiar L2 words was easier, as indicated by better recognition and cued-recall performance throughout the learning phase and in delayed tests. Furthermore, new-meaning learning facilitated, rather than impeded, the processing of original meanings, especially after a delay. Comparing these findings with those from previous studies involving native speakers, it appears that lexical quality influences how new and prior knowledge interact during new-meaning learning.