Pus ‘ta cabrón : variation of pues in the Spanish of southern San Diego

This dissertation documents linguistic variation in the Spanish of the border community of Southern San Diego. A diverse community that comprises recent migrants, established immigrant communities, and heritage speakers, it is an ideal area for linguistic inquiry. This work explores the discourse marker pues, which is ubiquitous in Mexican Spanish, where it is variously translated as 'well', 'so', 'since', and 'then' and presents a wide range of variation in its production. The present study focuses on a four-way distinction, contrasting standard pues [pwes] and three forms that carry social stigma: the archaic pos [pos], the vowel-raised form pus [pus], and the devoiced p's [ps]. The study entertained the following questions: (i) Do the various forms of the discourse marker pues in the Spanish of Southern San Diego differ in frequency of occurrence? and (ii) What are the linguistic (internal) and social (external) factors that condition the various realizations of pues in the Spanish of Southern San Diego? I sought to answer these questions following a quantitative variationist methodology. For that purpose, I analyzed forty-seven (47) sociolinguistic interviews collected between January and June 2022. These interviews were transcribed using Automatic Speech Recognition technology and the resulting transcripts yielded 4,431 tokens of the discourse marker. The study examined the variation in the production of the discourse marker pues as [pwes], [pos], [pus], and [ps] by considering a series of linguistic and social variables. The independent variables: position in the utterance and pragmatic function as internal variables, and gender, generation, social class, and ethno-racial identity as external variables. Statistical analysis of the data was carried out in R. The packages used for random forest and conditional inference trees was party, and the packages ggplot and VGAM were used for visualization and multinomial regression as required. The results show that the variant with the highest frequency in production was [pwes] at 44.93%, closely followed by [pus] at 37.26%. The data further reveal that this variation is conditioned by the social variables of gender, class, and self-reported ethno-racial identity. In particular, gender was a strong indicator for variation, as female speakers overall tended to default to the the standard [pwes], whereas male speakers produced [pus] at higher rates. However, socioeconomic class was seen to condition this variation, with women from the higher class also favoring pus over pues. This dissertation project offers insight into linguistic variation and vitality in an understudied region in Southern California as and contributes to growing research on the sociolinguistics of the borderlands.