On the Mental Lexicon : At the Linguistic/Psycholinguistic Interface.

点击翻译按钮获取中文摘要

The aim of this project has been to investigate the mental/neurological mechanism involved in morphological processes, taking into consideration the results reported in psycholinguistic/neurolinguistic studies.Specifically, we examined in a joint research whether the "dual mechanism model", proposed by S.Pinker and his colleagues for inflectional morphology, is appropriate for derivational morphology. The model argues for qualitative diferences between rule-governed regular processes and associative memory/analogy-based irregular processes. We conducted two experiments, one on normal adults and the other on brain-damaged patients, where we examined differences between two Japanese nominal suffixes -sa and -mi. The result of the first experiment shows that -sa and -mi behave differently with respect to applicability to novel words and that -mi, but not -sa, exhibits the "similarity effect", demonstrating that both rule and analogy are operative in derivational morphology. Furthermore, p … More atients with a lesion in Broca's area preferred -mi over -sa on novel words while patients with a lesion in other areas exhibited a different preference pattern, which suggests that the two mental mechanisms (rule and analogy) are supported by different neurological mechanisms.These findings have significant implications both for theoretical morphology and neurolinguistics : they provide a new perspective to the theory of morphology in showing that the rule/analogy distinction crosscuts the inflection/derivation distinction, and to the study of language and brain in demonstrating that Broca's area, which is believed to be involved in "syntactic" processes, is in fact responsible for "computational" processes whether they involve phrasal/sentential-level units or morphological/word-level units. Thus our study illustrates how researches in the two fields can benefit from each other.Our study has so far not given enough considerastions on how existing words with -sal/-mi are processed, and we will continue working on this with a new series of experiments. Less