Search

This thesis is an attempt to probe into the organisation of the L2 mental lexicon through sorting tasks, which Haastrup and Henriksen (2000) pioneered. Particularly, it addresses whether the lexicons of native speakers of English and non-native speakers of English are different from each other quantitatively and qualitatively. After a replication of Haastrup and Henriksen, five experiments each using a different set of 50 high frequency English words taken from different parts of speech were conducted on 30 participants in each group (28 for the first experiment). The studies found that LI and L2 differences were generally subtler than had been expected, such as for mean cluster number, cluster size, variability as well as the mean individual dendrogram and group dendrogram distances. However, cluster analysis showed that L2 lexical organisation was consistently different from LI lexical organisation. Thus, it is highly plausible that the L2 mental lexicon has developed lexical networks that are on the surface similar to the LI mental lexicon, when in fact the two lexicons have really developed different organisational structures from each other. Meanwhile, it was revealed that nouns can be predictors of LI and L2 differences in all the tested variables. This result was attributed to the fact that the Japanese language has a significant number of loanword nouns that have originated from the English language. This suggests that these L2 lexical items, which are first learned as LI lexical items as false friends, can be extremely difficult for Japanese L2 learners to re-leam and restructure into native-like L2 knowledge and organisation.