Variation in Weinreich Herzog and Labov's influential 1968 paper on the empirical foundations for at theory of language change became the primary weapon against the structuralist conception of change. It thus became inextricably bound to the concept of change. The focus of this paper is to untie this knot so that variation may be seen as a dimension of language in its own right. We do not necessarily have change whenever we have variation and conversely we do not necessarily have variation whenever we find change. The primary evidence for this claim will be an analysis of the a-variants of modern Danish viewed as more or less stable sociolinguistic variation.

In addition, I would like to diversify the notions needed to account for variation so that we agree on what we are talking about when we talk about variation. This makes a revisit to the famous dichotomies of competence and performance mandatory. Finally, I would like to raise the question of how to account for variation: rules? And if so: What kind of rules? How do we write grammars which incorporate variation?