Abstract
In an article published in 1981 together with Brigitte Schlieben-Lange, Harald Weydt emphasized the compatibility of structuralist methods and theory with facts of linguistic variation. The present paper tries to demonstrate, based on some remarks on linguistic history, how a structuralist view on linguistic variation is adequate, but limited to certain aspects. The paper tries to show that differences beyond linguistic structures can be as important as structural differences for speakers when they are trying to classify linguistic varieties. Furthermore, a combination of a description of phonological processes, common in generative phonology, and of a structuralist view is proposed, claiming that both are not contradictory, but compatible.