UC Davis

This dissertation examines the process of lenition, or spirantization, of intervocalic voiced stops /b, d, ɡ/ in heritage Spanish. While spirantization is one of the more extensively studied phonetic and phonological phenomena in Spanish, this study introduces methodological and theoretical advances for the study of heritage phonology. The data come from an understudied variety of U.S. Spanish spoken by heritage speakers of Colombian descent in New Jersey and Florida across three generations.

Previous studies have analyzed linguistic factors such as phoneme, stress, word position, and task type, that mediate voiced stop lenition in monolingual varieties of Spanish (Broś et al., 2021; Carrasco, 2008; Carrasco et al., 2012; Colatoni & Marinescu, 2010; Eddington, 2011; Figueroa, 2016), heritage Spanish (Amengual, 2019; Blair & Lease, 2021; Rao, 2014, 2015), and L2 Spanish (Cabrelli Amaro, 2017; Zampini, 1997, 1998). The present study builds on this research by expanding the variable of word position beyond the binary medial vs. initial distinction to include three prosodic domains: syllable-initial within the word, initial in the minimal prosodic word, and initial in the maximal prosodic word. The results from a reading task and conversation dyad confirm that the linguistic variables of phoneme, stress, prosodic domain, and task type mediate categorical acoustic measures of lenition in heritage Spanish in a way that replicates previous studies’ findings for other varieties. Social factors such as generation, order of language acquisition, and speech community are not found to be significant predictors of voiced stop lenition in the corpus of data analyzed here.

Based on the results of the experimental tasks and previous studies on other varieties of Spanish, this dissertation offers a formal analysis in phonetically based Optimality Theory that accounts for domain-medial spirantization as continuity lenition, which, in conjunction with fortition at domain edges, helps demarcate prosodic constituents for the listener’s benefit (Katz, 2016). The analysis makes testable predictions about possible patterns of variation both in and beyond Colombian heritage Spanish. In sum, this dissertation contributes to a growing body of work (Broś et al., 2021; Harper, 2014; Katz, 2016; Katz & Fricke, 2018; Katz & Moore, 2021; Keating, 2006; Kingston, 2008) seeking to flip the traditional narrative that consonantal weakening is primarily a speaker-based phenomenon driven by effort minimization.