Reading disorders in a language with shallow orthography: A multiple single-case study in Italian

Background: This study aimed (i) to verify whether the classical word-naming models developed for English-speaking participants also account for the performance of patients who speak a shallow-orthography language such as Italian, and (ii) to study the effects of word frequency, concreteness, and grammatical class on word naming. Methods & Procedures: A total of 90 Italian aphasic patients participated in two reading tasks. The first task contained four sets of items: (i) concrete nouns (natural objects and artefacts), (ii) abstract nouns, (iii) function words, (iv) morphologically simple legal nonwords. The second task (trisyllabic words with unpredictable stress position) was designed to test reading ability along the lexical route (the position of the major word stress is the only opaque variable in the Italian reading system). The patients' performances on the two tasks were analysed for strong dissociations, and to test the effect of grammatical class, concreteness, word frequency, and item length. The effect of age of acquisition was tested in a subsequent analysis. Outcomes & Results: Reading scores were pathological for all patients. The present sample reflected the entire spectrum of reading impairments: phonological (49), surface (4), undifferentiated (32), and letter-by-letter (5) dyslexia, which is in line with data reported for English-speaking aphasic patients. Only one of the phonological dyslexic patients made semantic errors (a reading impairment compatible with the diagnosis of deep dyslexia). The vast majority of Broca's aphasic patients suffered from phonological dyslexia (76%), while fluent aphasic patients were distributed more evenly across dyslexia types; all four surface dyslexic patients belonged to the fluent aphasia group. Overall, grammatical class (concrete nouns vs function words) had a significant effect on 14 patients (15.6%), concreteness (concrete vs abstract nouns) on 15 (16.7%), and word frequency on 5 (5.6%). Grammatical class and concreteness affected the performance of phonological and undifferentiated dyslexic patients, and seemed not to influence the scores of the surface dyslexic patients. Age of acquisition turned out to have a highly significant effect and may account for most of the lexical effects emerging from the first analysis. Conclusions: The entire spectrum of reading impairments was observed in this group of Italian aphasic patients, thus confirming the validity of contemporary reading models also for shallow-orthography languages. Concreteness and grammatical class effects, present in deep dyslexia, also affected the performance of patients suffering from other types of dyslexia, although both phenomena might derive from a confounding effect of age of acquisition.