Some Reflections on Italian Derivation and Computational Morphology

As known, Computational and Automatic Derivative Morphology routines are useful especially for speeding up the creation of exhaustive electronic dictionaries necessary to achieve complete and in-depth Natural Language Processing analyses. Furthermore, the structuring of these routines can be based on precise morphosyntactic and derivational rules, such as, for example, in Italian, the one that predicts, for each transitive verb, an adjectival derivation in –bile. However, it is worth noting that the Italian lexicon - together with many other lexicons - is not homogenously structured: this means thatits entries, as for their morphosyntax and semantic uses, are not all subject to the same rules of morphological derivation and segmentation. For this reason, a “wild card” application of Computational and Automatic Derivative Morphology routines, although well structured, is not possible.