II United States
4 Describing and analyzing variability in Spanish /s/
Describing and analyzing variability in Spanish /s/: A case study of Caribbeans in Boston and New York City
Daniel Gerard Erker and Madeline Reffel
Systematic variation in speech has proven to be a rich source of insight into human language. In particular, patterns of phonetic variability have illuminated the sounds of speech as physical phenomena produced by human vocal tracts, as psychological phenomena stored in human minds, and as social phenomena distributed across and within human societies (Ladefoged 1980; Ohala 1989; Foulkes & Docherty 2006: among others). With respect to phonetic variation in the Spanish-speaking world, perhaps no speech sound has figured more prominently than Isl, the voiceless alveolar fricative. In addition to its importance in major diachronic shifts in the organization of Spanish phonology (see Nunez, this volume), /s/ also played an early and important role in the emergence and recognition of patterns of social and regional linguistic variation within the Iberian peninsula. As early as the end of the fifteenth century, /s/ already constituted a site of synchronic variability, manifesting in “aspiration... ca'ida total... у ultracorreciones” “aspiration... complete loss... and hyper-correction” (Lapesa 1942: 387). Later, as Iberian ways of speaking spread across the Atlantic, the distribution of the phonetic variants of Isl became an important tool for philologists and dialectologists interested in tracking the development of Spanish in the Americas (Marden 1896; Urena 1921; Granda 1994; Hammond 2001).
Such dialectology initiatives subsequently informed more contemporary approaches to sociophonetic variation in the Spanish of the Americas (Resnick 1975; Canfield 1981), which in turn paved the way for deeper study of Spanish Is/ in its social context as well as further investigation of the linguistic factors that condition its realization. Indeed, a host of studies carried out in the tradition of variationist sociolinguistics have focused on Is/, resulting not only in an increased understanding of the phenomenon itself, but also in the refinement of theories of language variation and change (Cedergren 1974; Poplack 1980: Hochberg 1986: Cepeda 1995; Erker 2010; File-Muriel & Brown 2011; Chappell 2014). In other domains of contemporary linguistic inquiry, Spanish Is/ variation has played a number of important roles, featuring prominently in the development of exemplar-theoretic models of linguistic representation (Bybee 2000, 2002; Fox 2006), in furthering our understanding of the dynamics of sociolinguistic perception of speaker identity (Walker et al. 2014), and in accounting for the outcomes of language and dialectal contact (Lamboy 2004; Aaron & Hernandez 2007; Lynch 2009: Waltermire 2011; Erker & Otheguy 2016).
On account of this long and rich research tradition, contemporary scholars are well-situated to pursue further study of Spanish Is/ from a position of substantial understanding. However, this enviable vantage point is not optimal in all respects. It requires researchers to navigate an extant literature that is vast, with relevant studies easily numbering in the hundreds if not thousands (this is to say nothing of the extensive crosslinguistic research on the phonetics of fricatives more generally). More problematic than its sheer volume, however, is the wide-ranging descriptive and methodological variation present in the research on Spanish Is/. This is due, in large part, to the time depth of inquiry into /s/, which extends from our present smartphone era back to the dawn of the printing press. Lapesa’s observations dating systematic Is/ variation to at least as far back as the last decade of the fifteenth century are based on a letter written by the son of Christopher Columbus. While Urena and other pioneering dialectologists had shifted their attention away from written data and towards speech, there are, of course and unfortunately, no audio recordings of the people whose voices served as the basis for the earliest attempts at the dialectal zonification of Latin America. The relatively recent emergence of low-cost audio recording technology that coincided with the development of variationist sociolinguistics radically improved the quality and reliability of observations of variability in speech, a development whose benefits extended to the study of Spanish Is/. However, because sprectrographic or other means of detailed acoustic-phonetic analysis remained unwieldy and expensive, most of the early variationist work on Spanish Is/ relied on descriptions that were based exclusively on analysts’ perceptions of segmentally coded alternations. In the last couple of decades, as it has became possible to extract fine-grained acoustic measurements of the sounds of speech from any laptop, shifts in the best practices of sociophoneticians have led scholars of Spanish Isl to explore its spectrotemporal properties from a gradient, subsegmental perspective. But technological advance comes with its own problems, namely, a potentially overwhelming array of acoustic measures to choose from when describing Is/.
How then to best synthesize the methodologically diverse literature, especially in the context of carrying out future research? Specifically, how can we capitalize on existing studies to operationalize a method for describing and analyzing Is/ that is robust, reliable, revealing and reproducible? Answering these questions is the first task of this chapter, one that we will approach methodically, empirically motivating each aspect of the descriptive protocol that we settle upon. In the end, it will consist of a combination of measures, some of which are discrete and others that are continuous, some of which are measured using acoustic-phonetic software and others that are based on perceptual impressions.
Once the proposed protocol is laid out in full, we will take it out for a test drive, illustrating how it can be used to explore a research question in the context of a case study. Specifically, we will ask whether patterns of Is/ variation among Spanish speakers of Caribbean origin are either intergenerationally stable or undergoing change in a setting characterized by dialectal and language contact. The appropriate motivation and context for this question will be provided in due course. For now, it suffices to say that we will see our descriptive protocol (configured to reflect the primary insights of the research literature on Spanish /s/ as well as current understanding of the acoustic- phonetic properties of fricatives) does a very good job of capturing the kind of data needed to seriously engage the research question. It also leads to a clear answer, namely, that patterns of is/ variation present in the speech of recent arrivals from the Caribbean to two US cities, New York City and Boston, are conserved both among immigrants who have spent many years living in the United States as well as among US-born Spanish speakers of Caribbean ancestry. In other words, results reveal a trend of intergenerational structural continuity in Is/ variation among the study participants. As a first stop on the way to this conclusion, let us ask a simple question: What kind of phenomenon is Spanish Is/ variation?