Logo: to the web site of Uppsala University

The article takes as a starting point the claim by interactional linguists that speakers’ intraindividual variation within an individual conversation is best described and explained with reference to the local context of the utterance. Based on accommodation theory they argue that a good deal of a speaker’s conversation-internal variation can be seen as convergence with or divergence from his/her interlocutor and that this tendency is an important explanation of language change. This view is here called in question, supported by a comparison between instances of accommodation and grammatical and prosodic features in a large corpus of spontaneous speech by socially diversified informants. Not only do cases of divergence outnumber those of convergence but turn-internal variation, which cannot be interpreted as accommodation is more numerous than the sum of accommodations. The distribution of prosodic and grammatical variants in the three analyzed variables is mostly statistically significant. The results suggest that language-internal factors are at least as important as interactional ones for understanding variation. Some methodological and theoretical problems about accommodation theory are discussed, and considering these together with the empirical results, the article concludes that, along with social prestige, language-internal factors are somewhat better predictors of language change than accommodation.