Lettura di parole a morfologia complessa: un'analisi degli errori in bambini con e senza difficoltà di lettura

[Autom. eng. transl.] Introduction. Many studies have highlighted the process of morphological decomposition in tasks of visual recognition of complex words and pseudoparoles, both in adults and in children. In these studies, latency and accuracy of the response were analyzed, but the coding of the types of errors observed in reading aloud tasks could provide further elements for a more detailed description of the morphological effects. Given that we have a sufficient body of errors, and the accuracy of adults reading aloud is very high, we have considered a sample of children with and without reading difficulties. On the errors produced by them, a classification focused on the morphological structure (Morphology Coding Scheme, MCS) was developed, to verify the use of morphological parsing also in the production of errors. Participants. 54 primary school children (4th-5th class): 18 with reading difficulties and 36 norm readers. Materials and procedure. 71 derived names, 42 from nominal basis (eg artist) and 29 from verbal base (eg punishment), and 99 non-derivative names (eg condition) were presented in a task of reading aloud. Results. The coding of the errors through the MCS obtained an adequate inter-rater agreement (Cohen's k = .78). Most of the errors produced by children respect the morphemic structure of the target: in 70% of cases the morphemic constituents are replaced by other morphemes (bases with bases; suffixes with suffixes). Children with reading difficulties tend to less preserve the morphological structure of verbal-based names than nominal-based names. Conclusions. The data obtained with MCS confirm that children use morphemes as a unit of reading and support this type of analysis. The application of MCS to corpora of patients with developmental and acquired language disorders may provide indications regarding the impairment of morphosyntactic components in different pathologies.