One thing leads to another?

thanks to Li Wei for encouraging us to do this special issue, to Carol Pfaff and Tonjes Veenstra for organizing a Berlin workshop that allowed some of this article to be tested on an audience, and to all the colleagues mentioned in this article for being so inspiring. Special thanks to Fred Field and Helena Halmari for commenting on earlier versions of this article. Any errors are, of course, my responsibility. This article introduces the topic of and the contributions to this Special Issue of the International Journal of Bilingualism. It explores the degree to which the hypothesis that codeswitching is a cause of contact-induced language change makes sense. After reviewing a number of methodological conditions that need to be met before the question can even be tackled, I provide an overview of theories proposed to account for structural change in contact situations (Croft, Johanson, Thomason), pointing out commonalities and differences. The article concludes with an attempt to classify attested contact-induced changes on the basis of these theories, and finally revisits the question to what degree codeswitching can cause change. 1Introduction This special issue of the International Journal of Bilingualism is about the relationship between codeswitching and language change. Both terms are used in many different ways, so I will first explain what I mean by them. “Codeswitching, ” here, refers to any kind of discourse in which words originating in two different language systems are used side-by-side. “Language change ” should be read as ‘contact-induced structural change’, that is: changes in the structure of a language as a result of language contact. One may wonder why we should be interested in the connection between the two phenomena. This may be best illustrated through an example. Consider the following Australian German sentences, from Clyne (1967).1 1 See Clyne (1967, pp.64 – 68) for many other examples, as well as Clyne (2003)