Abstract

The presentation has a methodological focus and is illustrated with a few cases related to urban geography (education, urban growth, polycentrism). Modeling activity implies a close interaction between three domains: - the empirical one where the phenomenon of interest is observed and questioned; - the conceptual one where potentially explanatory relations are specified, based on a theoretical background; - the domain of the formal modeling where the model is developed and implemented. In an ideal world these three domains are perfectly compatible. However, in our everyday research practice, it is not uncommon to be confronted with gaps, for instance between concepts and “observables”. Most often we solve it with proxies. An ontological approach helps to identify such gaps and to value their effects on the modeling process. First I will present the three domains stated above and illustrate possible gaps with a case study on educational and social inequalities in space. I will then apply a grid of analyses based on the relations between these three domains on a series of cases corresponding to different kinds of gaps between concepts and “observables”.