Lexical Frequency and the Realization of Italian Dental Affricates

This study investigates the phonetic variation in Italian dental affricates, focusing on the role of lexical frequency, phonological context, geographical origin, and speech style (read speech vs. collaborative dialogue). Previous research on lexically based phonetics has emphasised the link between item frequency and phonetic realisation; in this paper, we test these premises on a class of rare and marked sounds—dental affricates. A corpus of read sentences and map-task dialogues produced by northern and southern Italian speakers was analysed acoustically with respect to voicing and duration. Results show that phonological context and geographical origin are the primary determinants of voicing, with Southern speakers favouring voiced realisations and word-initial position strongly conditioning voicing patterns. Lexical frequency does not significantly predict voicing category once between-speaker and between-item variability are appropriately modelled, but it does exert a positive effect on affricate duration: higher-frequency words contain relatively longer affricates, reflecting compensatory reservation of segmental identity within otherwise reduced words. Speech style significantly affects duration, with reading favouring longer realisations. These findings reveal a dissociation between categorical and gradient levels of phonetic variation, supporting usage-based models in which lexical frequency modulates fine-grained phonetic implementation rather than determining phonological categorisation.