This experimental laboratory-based study investigated the role of input modality in remembering name-referent associations in learning nonsense words. Three groups of L2 learners of English attempted to learn and remember name-referent associations in three different conditions: auditory only (n = 26), visual only (n = 28) and dual-modality (auditory/visual) (n = 25). Immediate recall and recognition of name-referent associations revealed no significant differences between the visual and auditory conditions. However, both recall and recognition of paired associations improved significantly when dual modality was used. These results do not seem to support the superiority of one mode of presentation over another for remembering name-referent associations in vocabulary acquisition. However, they do seem to provide support for the dual-modality hypothesis in vocabulary learning and the idea that semantic representations of words benefit from referential connections in which both auditory and visually-based processing is involved.