Variation theory and second language learning: Linguistic and social constraints on interlanguage tense marking

Interlanguage variation has become a central concern of second language acquisition research as scholars have come to realize that understanding variation is the key to understanding acquisition. Despite widespread recognition of the importance of variation, few researchers have explored the implications of the Labovian quantitative paradigm for theories of second language acquisition or used techniques of variable rule analysis developed by sociolinguists to model variation in native languages. This study, based on approximately 5,000 past-reference verbs and 3,000 final consonant clusters extracted from tape-recorded sociolinguistic interviews with twenty adult Mandarin-speaking learners of English, shows that variation in interlanguage tense marking and t/d deletion is highly systematic and subject to multiple linguistic, social, and developmental factors. Past tense marking is significantly affected by the salience of the difference between present and past tense forms, grammatical aspect, the speech situation, and English proficiency. In addition, marking of regular non-syllabics is constrained by the phonetic features of the preceding and following segments.;The effect of the convergence of two variable processes, t/d deletion and past tense marking, is also examined. Variable rule analysis confirms the hypothesis that convergence has an additive effect, resulting in a greater likelihood of t/d absence from non-syllabic preterits than from monomorphemes. A comparison of t/d deletion in native and nonnative speech suggests that to replicate native-speaker performance, nonnative speakers must acquire a system of constraints that limit deletion across internal morphological boundaries as well as rules governing past tense marking and the use of past participles.;The quantitative results suggest that acquisition of phonological and morphological rules proceeds along a multi-dimensional continuum, with linguistic, social, and developmental factor groups constituting the different dimensions. The results, which offer strong confirmation of the variable rule hypothesis, also suggest that the speech of language learners is an appropriate and too little explored testing ground for sociolinguistic theories of variation.