Communications of the Association for Information Systems


Discussions of conceptual modelling assume that notation details are of secondary importance, a matter of taste and past experience rather than of science. For example, it does not really matter if cardinality is shown with a 'crow's foot' or with the symbol '1..*'. This paper argues that such an assumption is wrong and that the notation is extremely important in the process of modelling relationships because of the normative language that the notation specifies. Normative language is shown to be a useful way of understanding and comparing relationship notation. Barker's practical relationship definition, using a formal notation, is shown to be sufficiently well formed to allow the modeller to make sense of the domain in their own linguistic context. Less formal notations are shown to disadvantage the less experienced modeller