Learning the meaning of new words: Behavioural and neuroimaging evidence

This thesis aimed to shed light on the process of word learning and the consequences of storing, retrieving, and using new lexical representations. A number of behavioural experiments and one final fMRI study were conducted. Experiments 1-2 investigated effects of context variability (number of different contexts) and semantic richness (number and type of semantic features) on word learning in a second language. Experiment 1 suggested that context variability benefits word naming and semantic decision. Experiment 2 showed that semantic richness leads to better performance in semantic decision and cued recall, but does not affect word naming or recognition memory. Experiment 3 investigated semantic richness effects across speakers of English L1 and English L2. Results showed that participants did not differ regarding recognition memory and semantic decision; however, L1 speakers outperformed L2 speakers in naming and cued recall. Experiments 4-5 investigated the time course of word learning and examined effects of semantic richness at two different time points. The findings suggested that semantics affects recognition memory, but only a week after training. Effects of semantic richness on categorization and cued recall were found a day and a week after training, with participants showing improvement over time in all conditions. Experiments 6-7 assessed whether improvement over time in Experiments 4-5 was simply due to the passing of time or due to the effect of previous test instances. Results showed that performance improved over time in all tasks when participants were tested on the same novel words, but declined when they were tested on a different set of novel words, suggesting that performance only improved if mediated by a test instance. Experiment 8 was aimed at collecting semantic features from British speakers for 100 familiar words. Finally, Experiment 9 explored the neural correlates of familiarity and semantic richness. Two distinctive brain networks were observed during the categorization of familiar and novel words, consistent with previous findings. Rich semantics was associated with increased activation in conceptual representation areas, whereas poor semantics was reflected in heightened response in semantic control areas. The findings of this thesis have important implications for theories of word learning and semantic memory.