The farmer, the pioneer and the corner: a tale about derived word processing

Two main approaches have been proposed to describe the production of morphologically complex word: some models conceive a direct access to full-form representations from the semantic system, while other architectures posit a lemma layer between the semantic system and a morpheme-based lexicon. In this study these hypotheses are compared by assessing the reading performance of a patient, GR, affected by deep dyslexia with reading impairment of non-words, function words and verbs, and semantic and morphological errors. We first tested the role of semantic properties in modulating morpheme access. 138 Italian nouns were employed as stimuli in a reading task: 43 derived nouns in which the root and the suffix have a transparent relation (as in the English noun “farmer”), 49 derived nouns in which the root and the suffix have an opaque relation (as in the English noun “corner”), and 46 pseudo-suffixed nouns with non-existent pseudo-roots (as in the English noun “pioneer”). Roots were frequently produced in both transparent and opaque stimuli; pseudo-suffixed words were equally impaired, but their pseudo-roots were never produced. In order to investigate the effect of the morpheme grammatical class, GR was also asked to read aloud a list of 80 Italian derived nouns, having either verb (“negazione”, negation, from “negare”, to deny) or noun (“occhiata”, glance, from “occhio”, eye) roots. GR’s performance was influenced by root grammatical classes: noun roots were reported more often than verb roots. Considered together, these results are in line with models positing (de)composition of derived words and a lemma-lexeme architecture.