A Sociophonetic Investigation of Unstressed Vowel Raising in the Spanish of a Rural Mexican Community

Jennifer Barajas, Ph.D. (currently at Bradley University)

While the vocalic system in Spanish is typically described as stable, there is increasing evidence for dialectal vowel variation. The focus of the current study is variable vowel raising in rural Michoacan Spanish, which involves the mid vowels, /e/ and /o/, in unstressed post-tonic positions and their variable realizations as [e, i] or [o, u], respectively. In this dissertation, I perform a sociophonetic study of this variable process using acoustic information about the formant values of the relevant vowels and I examine the influence of different linguistic and social factors on this phenomenon. Previous studies on unstressed vowel raising mainly focus on Puerto Rico (Holmquist 1998, 2005 and Oliver Rajan 2007, 2008). Several studies mention that vowel raising occurs in Mexico, but none offer a thorough description of the process. My dissertation focuses on the variation in mid vowel realizations found in the Spanish spoken in Colongo, Michoacan, Mexico, and it is the first systematic description of vowel raising in the region.

As for the independent variables, I look at linguistic factors that have been shown to influence raising in other dialects such as the tonic vowel, preceding consonant, following sound, stress pattern, and whether the vowel is in an open or closed syllable, among others. The analysis of these factors allows us to examine the relation between vowel raising and another vocalic process present in Mexican Spanish, namely unstressed vowel devoicing (UVD), by which unstressed vowels become devoiced or reduced in their duration. In this dissertation, I propose that both of these processes are cases of vowel weakening and I provide evidence to support this claim. Given this Weakening Hypothesis, I expect vowel raising to occur in the same, weak positions where UVD occurs, namely in closed syllables and utterance-finally, and my results support this idea. As part of the Weakening Hypothesis, I also propose that weakening will lead to an increase in coarticulatory effects, which are more likely in contexts where there is reduction in duration (Browman and Goldstein 1989, 1992). Thus, I expect to see these coarticulatory effects from both preceding and following sounds in addition to effects from the tonic vowel, and my results find this to be true as well.

The social factors of age, education level, occupation, and mobility, since they are highly correlated, are used in an analysis of social networks to look at social differences within the community. I create a numerical index incorporating these factors that ultimately results in the division between an open or closed social network. I expect less vowel raising in the open social network, where participants have more access to interactions outside of Colongo, and more vowel raising in the closed social network, where participants have less outside influence, and my results confirm this hypothesis. The division into social networks helps us to better understand the social differences among the community members and the effect of these differences on vowel production.

Available at https://etd.ohiolink.edu/ap/10?0::NO:10:P10_ACCESSION_NUM:osu1403808807