Persistence in Spoken English: A Variationist Approach

Speakers go for recently used or heard linguistic options whenever they can; this tendency is referred to as persistence in the present paper. Although discourse analysts have long known about the important role repetitiveness plays in discourse (for instance, Tannen 1987) and although there is a sizable body of psycholinguistic research on ‘production priming ’ (for instance, Bock 1986), the issue has received less attention in variationist sociolinguistics than it deserves. The present paper seeks to demonstrate that persistence can be neatly operationalized in variationist research designs. Drawing on data from several corpora of spoken English and utilizing binary logistic regression, the paper investigates, as case studies, two well-known alternations in the grammar of English – future marker choice (John will see Mary vs. John is going to see Mary) and complementation strategy choice (John starts to wonder vs. John starts wondering) – with regard to the explanatory power yielded by persistence. The paper presents evidence that the corpus data dovetail nicely with psycholinguistic findings on priming, and that variationist models omitting persistence would leave a substantial share of the observable variation unaccounted for, or even erroneously identify it as ‘free ’ variation although it is clearly patterned