In this chapter I am concerned with explanation in the theory of linguistic change (on which see Lass, 1980).1 Within the broad context of explanation, I am more immediately concerned with the kind of interpretations that are placed on the variable data that we derive from empirical investigations in speech communities. Thus, the chapter is not primarily data-based, but is mainly about argumentation in sociolinguistics. In particular, I shall focus on the compatibility (or lack of it) between interpretations that are dependent on the notion of speaker-identity (particularly as this is accessed through the social network model) and interpretations that depend on the notion of prestige.