The world in which we live is filled with sensory experiences. Language provides us with a manner in which to communicate these experiences with one another. In order to partake in this communication, it is necessary to acquire labels for things we see, hear, feel, smell, and taste. Much is known about how we learn words for things we can see, but this bias in the literature leaves many open questions about words attributed to other modalities. This cumulative dissertation aims to close this gap by investigating how both 10- to-12-month old infants and adults map novel pseudowords onto environmental sounds in an auditory associative word learning task with the aim to explore how humans learn words for things that cannot be seen, such as thunder, siren, and, lullaby. Infants were found, via event-related potentials (ERPs), to be successful at auditory associative word learning, while adults are much stronger learners in multimodal audio-visual conditions. Across the lifespan, sensory modality was found to affect word learning differently in infants than in adults. Where infants benefitted from unimodal auditory word learning, adults were more successful in multimodal audiovisual paradigms. Furthermore, the modality of the object being labelled modulated the temporal onset and the topological distribution of the N400 ERP component of violated lexical-semantic expectation. Lastly, the temporal congruency of presented stimuli affected word learning in adults in an inverted manner to other forms of statistical learning. Word learning is sensitive to age, modality, and means of presentation, providing evidence for various intertwined learning mechanisms and bringing us a step closer towards understanding human linguistic cognition.