Enriched sublexical representations to access morphological structures. A psycho-computational account

The paper investigates the morphological impact of quantitative properties of lexical and sublexical structures in the decomposition of morphologically complex words by means of an activation-based simulation. A complex nucleus of blind-to-semantics relationships turns out to allow the emergence of proto-morphological representations and to provide a cognitively efficient route to complex word decomposition. A computational account of how this bundle of information is handled with in the lexical processing of complex pseudo-words is provided. The model is tested against an edit-distance algorithm and a non-word similarity rating task performed by a group of native speakers.