An Initial Examination of Southwest Spanish Vowels

Abstract

ABSTRACT. This article investigates the Spanish vowel system of Southwest Spanish speakers through an acoustic examination of F1 and F2. The corpus is based on a semi-spontaneous narrative by four female speakers. Repeated measurements of all five Spanish vowels in a stressed syllable are plotted, as well as a comparison of 30 productions of /a/ in a stressed and unstressed syllable. The findings indicate several shifts in the generally accepted Spanish vowel triangle including a lowering and fronting of /u/, a lowering of /o/, and a fronting of /a/ to the vowel space typically described for English /ae/. There was no reduction of unstressed /a/ tokens to a schwa. * ********** INTRODUCTION. while English dialectal variation of segments has typically focused on differences in vowel production, dialectal and sociolinguistic studies of Spanish segmental variation have tended to concentrate on consonants (Zamora and Guitart 1982, Lipski 1994). This focus on vowels reflects Navarro Tomas' (1977) claim that Spanish vowels are more steady and firm than consonants and dialectal variation affects Spanish consonants more so than vowels. It is precisely these claims of invariability that are examined in the research presented here. There are typically few studies of Spanish vowel variation, either of a historical nature as in the case of dialectal variation in the Spanish Peninsula, or of a synchronic perspective as motivated by language contact, e.g. in the case of indigenous languages. This study examines acoustically the Spanish vocalic system of fluent Southwest Spanish speakers (hereafter SWS) and compares the results with vowel system descriptions from seminal works on Spanish vowels. Specifically, the research presented in this study examines data from narrative speech and the formant values of the five contrastive vowels of SWS, as well as the role of lexical stress on the realization of /a/ in tonic and atonic positions. The organization of the article is as follows: Section 2 reviews the literature on acoustic characterizations of the Spanish vowel system. Section 3 details the methodology for the study, including a discussion of laboratory speech versus more natural elicitation techniques. In Section 4 I present the findings on the acoustic examination of the SWS vowel system. Section 5 contains a discussion of the findings with respect to vowel variation in Spanish dialects, the role of stress in the realization of SWS vowels, and the role of contact in the variation observed. The conclusions of the study are reviewed in Section 6. 1. Review of literature. The Spanish vowel system has five contrastive vowels. Since Navarro Tomas (1977), it has generally been understood and accepted that the Spanish vowel system is typically stable. Navarro Tomas (1977) suggests there may be slight differences between stressed and unstressed vowels, but that the differences are likely below a perceptual level, and in no way similar to the type of reduction experimented by Catalan. This claim has been echoed by others with respect to reduction to schwa of unstressed vowels in the English vocalic system (Stockwell & Bowen 1965, Quilis & Esgueva 1983, Barrutia & Schwegler 1993). 1.1. SPANISH VOWEL SPACE. In most characterizations, the shape of the Spanish vowel system/space is roughly triangular, exemplified in Figure 1. Generative accounts of Spanish /a/, utilizing binary features of height and frontness, differ slightly from more descriptive characterizations due to differing ends. For example, the Spanish low vowel /a/ is typically described as (+back) in generative analyses (Cressey 1978, Nunez-Cedeno, Morales-Front, Prieto I Vives, & Hualde 1999), while more descriptive characterizations portray /a/ as low and central (Whitley 2002:28, Barrutia & Schwegler 1993). [FIGURE 1 OMITTED] 1.2. ACOUSTIC ANALYSES. The few quantitative studies of Spanish vowels employing an acoustic formant analysis are based on laboratory speech, i. …

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 51 citations exist upstream.