ABSTRACT
This chapter introduces structural equivalence and reports an analysis of the Raleigh data that uses a metric of structural equivalence between nodes. The structural equivalence analysis contrasts with the structural cohesion analysis with respect to both the underlying question and the network metric. Four hypotheses guide the analysis; these concern the interrelated effects of structural equivalence, age, occupation, and sex on Southern vocalic features in Raleigh. In the present analysis, structural equivalence is calculated as Jaccard distance (called equivalence distance here), computed for each pair of speakers. Regression models using the Quadratic Assignment Procedure are used to evaluate the relationship between matrices of linguistic differences and matrices representing the social factors for each of the three generations in the Raleigh data. In Generation 1, the significant network effects are scarce and modest, reflecting the fact that the contact-induced retreat from the SVS had not yet begun. In Generation 2, significant main effects of equivalence distance emerge, as do interactions with occupation. In Generation 3, there are significant main effects of equivalence distance for /ɪ, e, ɛ/ as well as interactions among equivalence distance, occupation, and sex.
