Despite more than 40 years of code-switching (CS) research there is still considerable disagreement on some of the most basic notions on which competing theories depend.  In this talk I will attempt simultaneously to introduce beginners to the field and encourage seasoned CS researchers to questions some of their tacit assumptions. I will do this by focusing on conflicting meanings of the term ‘switch’, on disparate views as to the role of linguistic integration in CS, and on the little explored issue of the role of community norms in establishing CS patterns.  I will show how the understanding of switch depends on whether investigators prioritise linearity or hierarchy, and on how linguistic integration is assumed to help in distinguishing switches from borrowings. Finally, I will argue that neither the search for universal structural constraints on CS nor the more recent study of variation in CS have paid sufficient attention to the role of community norms and their acquisition by children.