In three experiments, we examine the effects of semantic context and word concreteness on the mapping of existing meanings to new words. We developed a new-word-learning paradigm in which participants were required to discover the meaning of a new-word form from a specific verbal context. The stimulus materials were manipulated according to word concreteness, context availability and semantic congruency across contexts. Overall, participants successfully learned the meaning of the new word whether it was a concrete or an abstract word. Concrete word meanings were discovered and learned faster than abstract word meanings even when matched on context availability. The present results are discussed considering the various hypotheses that have been used to try to explain the "concreteness effect". We conclude that the present investigation provides new evidence that the concreteness effect observed in learning is due to the different organization of abstract and concrete conceptual information in semantic memory.