We investigated the performance of two connectionist neural networks with different architectures to explore the degree of learning in generating the past participle form of Italian verbs on the basis of phonological characteristics. The networks were trained to generate the past participle form of verbs from different inflected input forms. We examined the degree of learning relative to the type of inflection given as input, the type of suffix produced, the classification of each verb according to the thematic vowel, the regularity of the stem and of the suffix. The networks were able to learn both regular and irregular forms, but the effect of regularity depended on the distributional properties of the conjugation to which a verb belongs, and on information provided by the input.
The role of phonology in the inflection of Italian verbs: A connectionist investigation
PASINI, Margherita;
2006-01-01
Abstract
We investigated the performance of two connectionist neural networks with different architectures to explore the degree of learning in generating the past participle form of Italian verbs on the basis of phonological characteristics. The networks were trained to generate the past participle form of verbs from different inflected input forms. We examined the degree of learning relative to the type of inflection given as input, the type of suffix produced, the classification of each verb according to the thematic vowel, the regularity of the stem and of the suffix. The networks were able to learn both regular and irregular forms, but the effect of regularity depended on the distributional properties of the conjugation to which a verb belongs, and on information provided by the input.
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
