ABSTRACT
We examine the degree of coherence in Heritage Italian speech to understand the degree of coherence in a community’s linguistic system. Coherence, in this context, is defined as covariation: in a coherent system, innovators in one variable will also be innovators in others, but extended to a language that has been less studied in sociolinguistics. Our findings counter claims from the experimental/acquisition literature that variation among heritage speakers stems from incomplete acquisition. The data are from conversational speech from homeland speakers (in Calabria, Italy), immigrants from Calabria to Toronto, and children of immigrants. Regression models were constructed for six dependent variables (prodrop, VOT (2 contexts), DOM, apocope, speech-rate). Speaker random effect estimates from the best-fitting models constitute a speaker list ranked by rate of use of the innovative form, controlling for the effects of uneven distributions across contexts. We calculate, pairwise, correlation coefficients across these ranked speaker lists. Rarely do the same speakers have the highest usage rates across the variables. We find little correlation to factors expected to affect heritage language speakers, as in a similar analysis of Heritage Cantonese.
