This article examines the claims of the 'dynamic' model of language variation by testing it using data recorded in northern French Avion. Ultimately these claims appear untestable in the absence of sound homogeneous data. The researcher is constrained to reject the model or the theory-data as "performance" misfit.