University of Pennsylvania Working Papers in Linguistics


The effect of ethnicity on sociolinguistic variation has been studied quantitatively from multiple angles. The existing methodological range has uncovered that ethnicity is reflected in various aspects of social life. However, the diversity of existing measures has i) prevented a direct comparison between studies and ii) relied on subjective quantification of social data. This paper introduces a novel methodology that unites previous measures and removes the need for subjective quantification. Specifically, I apply hierarchical cluster analysis to social data collected from an ethnographically-informed Polish-involvement survey to assess the impact of multiple social factors on regional variation for Polish New Yorkers. The analysis identifies the maintenance of transnational ties as the strongest predictor of linguistic variation and reveals speakers’ “hyphenated” (Polish-American) identities.