Second-Language Acquisition and Variationist Linguistics

The late 1960s and early 1970s witnessed the development of two subfields of linguistics: the quantitative study of linguistic variation, pioneered by Labov (1966), and the systematic study of second-language acquisition (SLA). Both areas shared a concern with understanding the underlying systems of nonstandard language varieties, often socially stigmatized varieties in the case of quantitative sociolinguistics. The variationist paradigm was soon extended to other socially stigmatized varieties spoken in cities around the world.

Several early SLA studies also focused on socially marginalized speakers, often immigrants to North America from developing countries. And, even in cases where SLA researchers studied privileged speakers, their concern was with discovering the underlying systematicity of variable learner production. This concern was a logical outgrowth of Selinker's (1972) original formulation of interlanguage as a learner's "approximate system," which shared features with the learner's first language and with the target language but was fully explainable by neither. Noyau (1990, 144-45) elaborated the idea of interlanguage and characterized the task of the SLA researcher as being "to describe . . . learner languages, which are to be considered as unknown languages." If learner varieties are "unknown languages," it follows that, like all human languages, they must be characterized by "structured heterogeneity" (Weinreich, Labov, and Herzog 1968, 99-100).

Despite the convergence of interests in understanding variability in language use and the speech of socially marginalized groups, variationist linguistics has had relatively little influence on SLA research. To be sure, several studies in the 1970s and early 1980s employed variable rule analysis (e.g., Dickerson 1974). However, until the late 1980s, variationist studies were rare in SLA. Preston (1996, 25) argues that the neglect in SLA research of the insights to be gained from variationist linguistics resulted from (1) the dominance of formal models in SLA, (2) the reduction of sociolinguistic aims to "socially sensitive pragmatics," and (3) misunderstandings by SLA researchers of basic concepts of the variationist enterprise. Here I shall discuss one consequence of the first factor--the overemphasis in SLA on the standard language--and suggest one area where variationist linguistics may make an important contribution to SLA research.

Most research in formal linguistics has focused on standard varieties. Within the formal tradition, researchers directly access the language via their own linguistic competence. Since very few academically trained linguists are native speakers of vernacular dialects, the reliance upon linguists' own intuitions necessitates a concentration on standard varieties (Milroy 1987). SLA researchers, of course, cannot rely on their own grammaticality judgments. Nevertheless, as in formal linguistics, the preponderance of SLA research has tended to equate the target language and the standard variety. Acquisition has often been judged on the basis of learners' supplying target language forms in obligatory contexts.

Consideration of the social structure of North American immigrant communities suggests that the assumption that the standard variety is the target language for all learners raises fundamental questions about what constitutes acquisition. Most immigrants belong to lower socioeconomic groups. Native speakers of the dominant language with whom they interact are more likely to be speakers of vernacular dialects than of standard varieties. In such circumstances, acquisition needs to be judged not in terms of the standard language, but in terms of the varieties with which learners are in most frequent contact. A simple example will illustrate the point.

It is well established that third-person singular -s is acquired late by learners of English (Goldschneider and DeKeyser forthcoming). It is also well established that third-person singular -s is highly variable in many vernacular dialects, among them African American Vernacular English (AAVE) (Wolfram and Schilling-Estes 1998). Puerto Rican migrants acquiring English in New York City often live in close proximity to African Americans, and many identify more closely with AAVE speakers than with speakers of the white middle-class standard (Zentella 1997). In such cases, we cannot assume that the absence of third-person singular -s represents a failure...