Code-switching is often studied in purely synchronic terms, as recorded speech is analyzed for patterns of language mixing. Though this has yielded numerous useful theoretical advances, it has also shielded the code-switching literature from serious engagement with the phenomenon of language change, even from the subtype of change caused by language contact. There is also the additional practice of limiting the study of code-mixing and code-switching to lexical mixing. On the other side of the fence, meanwhile, discussions of contact-induced language change tend to be limited to morphological and syntactic phenomena. This chapter breaks through this stalemate, and argues that a usage-based approach to language change actually demands integration of these perspectives. Code-switching should be seen as a reflection of lexical change. It is for this reason that a synchronic distinction between loanwords and code-switching makes no sense, since the terms refer to the diachronic and synchronic planes, respectively, of the same phenomenon. In the chapter, the author interprets the code-switching literature from this theoretical viewpoint, and explores what both the literature on code-switching and that on contact-induced change stand to gain from linking their empirical findings to a usage-based theory of language change that allocates proper attention to synchrony and diachrony, and unites lexical and structural change in the same framework.