Definizione di un modello computazionale della variazione dialettale basato sull'integrazione di fattori socio-demografici e geografici

In this study, we used a mixed-effects logistic regression model in combination with generalized additive logistic modeling to predict lexical differences in Tuscan dialects with respect to standard Italian. We used lexical information for 170 concepts in 213 locations in Tuscany. Although geographical position is an important predictor with locations distant from Florence having lexical forms more likely to differ from standard Italian, several other factors emerged as significant. The model predicts that lexical variants used by older speakers and in smaller as well as poorer communities are more likely to differ from standard Italian. The impact of the demographic variables, however, varied from concept to concept. For a majority of concepts, smaller and poorer communities have lexical forms different from standard Italian. For a smaller minority of concepts, however, larger and richer communities have lexical forms different from standard Italian. Similarly, the effect of speaker age and the average community age also varied per concept. While not significant as a fixed effect, the concept frequency showed significant geographical variation. These results clearly identify important factors involved in dialect variation at the lexical level. In addition, this study illustrates the usefulness of mixed-effects regression techniques together with generalized additive modeling for analyzing lexical dialect data.