How verb and noun bases can differently influence the ability to read aloud derived nouns: Evidence from typically developing children and children with dyslexia

This contribute is aimed at assessing whether the differences in verb and noun processing, found in several studies on children and adults with and without linguistic impairment, can lead to a different availability of noun and verb bases, when children try to apply their morphological knowledge to complex word reading. To verify if the grammatical class of the base word determines the role exerted by base frequency on reading aloud derived words, 18 children with dyslexia and 36 typically developing children, attending 4th and 5th grades of primary school, were asked to read aloud 52 nouns derived from noun bases (e.g., artista, artist), 43 nouns derived from verb bases (e.g., punizione, punishment), and 100 simple pseudoderived words (e.g., intervista, interview; condizione, condition). Mixed effect regression analyses were employed taking participants and items as random effects. As for latencies, the grammatical class and the frequency of the base did not influence the performance of typically developing children, while in children with dyslexia we observed a facilitatory effect of the base frequency, but only for noun bases. In the case of verb bases, no effect was found. Accuracy data show that, in both groups of children, the presence of a high frequency base leads to a higher level of accuracy and verb bases are associated to a higher error rate than noun bases. The role of base frequency confirms that in 4th and 5th graders morphological decomposition is likely to affect reading aloud long complex words, and the effect of grammatical class shows that base morphemes might bear with them grammatical features, arguably represented at a morpho-lexical level. The nature of the difference between noun and verb classes is still under debate, but the large amount of data supporting the role of grammatical class on the representation and processing of the language, more evident when the elaboration is not fully automatized, suggests that this variable is to be taken into consideration in assessing and modelling linguistic competence in a life-span perspective.