It appears you don't have support to open PDFs in this web browser. To view this file, Open with your PDF reader
Abstract
The present study investigates Spanish /b,d,g,s/ variation in the Rio Grande Valley, an understudied community in the Texas borderlands. This study seeks to fill a gap in previous literature by considering the role of identity and language ideologies in predicting /b,d,g,s/ variation. Previous research on Spanish /bdg/ realization finds that heritage speakers follow similar spirantization patterning to that of monolingual dialects of Spanish, where intervocalic /bdg/ is typically realized as an approximant [β̞ð̞ɣ̞]– rather than the voiced plosives [bdg] – as in <abuelo> ('grandparent') [a.'βu̯e.lo] (e.g., Amengual, 2019; Au et al., 2002; Blair & Lease, 2021; Rao, 2014, 2015; Regan, 2022). Though, some studies suggest that /bdg/ spirantization is more lenited in heritage Spanish, indicating greater reduction than in monolingual dialects (e.g., Blair & Lease, 2021; Rao, 2015).
Although /bdg/ variation is well-documented, research on /s/-voicing in heritage Spanish remains limited, with prior work focusing on anticipatory /s/-voicing assimilation, such as in <chisme> ('gossip') ['t͡ʃiz.me] where /s/ becomes voiced when followed by a voiced consonant (e.g., Boomershine & Stevens, 2021; Brown, 2020). While previous research emphasizes the effect of generation of immigration, with earlier generations aligning more closely to monolingual norms (e.g., Amengual, 2018; Blair & Lease, 2021; Henriksen, 2015), few studies explore how identity or Critical Language Awareness interact with linguistic variation of heritage Spanish.
Thirty bilingual speakers of Spanish and English from the Rio Grande Valley participated in an oral interview, Picture Naming Task, and background survey. The first 20 intervocalic /bdg/ and /s/ tokens after five minutes of the interview recordings as well as all intervocalic /b,d,g,s/ tokens from the Picture Naming Task were analyzed in Praat. For /bdg/ tokens, mean intensity difference of the approximant [β̞ð̞ɣ̞] segment was measured (the valley intensity of [β̞ð̞ɣ̞] divided by the average peak intensity of surrounding vowels). For /s/ tokens, voicing percentage was analyzed, measured as the percentage of the total segmental duration in which voicing was present.
Most intervocalic /bdg/ realizations were approximants in both the interview (70.4%) and PNT (79.0%). Elision also commonly occurred 24.4% of the time in the interviews and 11.3% of the time in the PNT. For /s/-voicing, most tokens indicated some percentage of voicing in both tasks (PNT = 68.2%; interviews = 76.8%) and less than a quarter of all tokens were produced as completely voiceless (Total = 24.8%; PNT = 31.8%; interviews = 23.2%). The results of regression analyses find that higher community identity and higher Critical Language Awareness levels predict more /s/-voicing and more /bdg/ elision. Conversely, higher ethnic identity related to less /bdg/ elision and less /s/-voicing. Altogether, these results suggest that a person who relates strongly with the RGV, who relates less strongly with being Mexican/Mexican-American, and who demonstrates greater acceptance of dialectal variation and translanguaging is a person more likely to exhibit /b,d,g,s/ reduction.
This investigation offers new exploration of methodological approaches to understanding language variation in a bilingual Spanish-English community by (1) providing the first exploration of intervocalic /s/-voicing in heritage Spanish; (2) using Lotería as a culturally-indexed PNT; (3) exploring the role of identity and Critical Language Awareness in predicting phonological reduction; and (4) the exploration of an un(der)studied borderlands dialect and inclusion of a non-university-affiliated social network. These methodological advances not only enhance our understanding of this community, but also open new directions for further study of variable phonetics/phonology in bilingual systems.





