Native and non-native processing of morphologically complex words in Italian

The present work focuses on the organization of the mental lexicon in native and non-native speakers and aims at investigating whether words are connected in the mind in terms of morphological criteria, i.e., through a network of associations establishing when a co-occurrence of form and meaning is found. Psycholinguistic research on native lexical access has demonstrated that morphology indeed underlies the organization of the mental lexicon, even though controversies about the locus of this level of organization remain. On the other hand, research in the field of second language acquisition has only recently turned to investigate such issues and its findings so far have been controversial. Specifically, the debate centers on whether native and non-native speakers share the same processing systems. According to recent proposals (Heyer & Clahsen 2015), this would not be the case and L2 processing would be more affected by formal rather than morphological criteria. In this light, the present work is aimed at verifying the impact of formal characteristics in native and non-native lexical access focusing on the processing of formally transparent versus non-transparent words in Italian. Two morphological phenomena are investigated by means of four psycholinguistic experiments involving a lexical decision task combined with the masked priming paradigm. Experiments 1 & 2 compare the processing of allomorphic vs non-allomorphic derivatives, to investigate whether formal alterations impair the appreciation of the relationship between two morphologically related words. Experiments 3 & 4 are focused on lack of base autonomy found in so-called bound stems, i.e., stems which cannot occur in isolation and are aimed at determining whether the processing of free and bound stems differs. The results of Experiments 1 and 2 indicate that allomorphic variation does not influence the associations established among related words in native speakers, in line with the predictions that can be formulated within usage-based perspectives on language. Non-native speakers, on the other hand, seem to be more pervasively affected by the phonological/orthographical properties of words, but not to the point that transparent morphological relations can be reduced to mere form overlap shared by morphological relatives. Likewise, stem autonomy was not found to affect the way words containing bound and free stems are processed by native speakers, at least under certain conditions, suggesting that boundedness is not an issue influencing the establishment of morphological relationships among words. Non-native speakers, however, were found to be sensitive to the isolability of the stem, in a way that suggests that free bases may be more salient morphological units for them, as opposed to bound stems, which are seemingly more closely associated with orthographic strings resembling each other. Taken together, the findings of the present work suggest a model of the native mental lexicon based on words and morphological schemas emerging from the relationships establishing among them, despite phonological variations and stem boundedness. While it is unclear whether such a system of connections and schemas is equally strong in the non-native lexicon, morphological relationships still appear to drive lexical organization. Crucially, however, such organization is modulated by form, as demonstrated by the effects of phonological variations and lack of base autonomy.