The mental representation of compound nouns: evidendence from neuro and psycholinguistic studies

There is a general debate as to whether constituent representations are accessed in compound processing, and which compound properties (e.g., headedness, semantic transparency) would influence this parsing procedure. This thesis investigates the mental representation of compound nouns in a series of six studies exploiting the properties of the Italian language, in the fields of both psycholinguistics and cognitive neuropsychology. First, effects related to the compound structure were investigated in the context of neglect dyslexia (Chapter 1). Second, converging evidence in favor of the headedness effect was sought in a constituent-priming experiment on normal participants (Chapter 2) and through the assessment of compound naming errors in patients suffering from aphasia (Chapter 3). Third, the access to grammatical properties of the constituents was studied in a single case study on deep dyslexia (Chapter 4). Fourth, the role of compound semantic transparency was investigated by assessing constituent frequency effects in both lexical decision latencies (Chapter 5) and fixation durations during compound-word reading (Chapter 6). The results indicate that the variables related to the whole compound (i.e., compound headedness, whole-word frequency and semantic transparency) play a crucial role in word processing, but also that constituent representations are accessed. To explain the observed effects a model will be proposed, positing both a multiple-lemma representation of compound words and a parallel procedure dedicated to the conceptual combination of compound constituents.