Scaling laws, when applied to geographical entities, reveal the configuration of the dynamic processes that generate inequalities in dimension. Two interpretations of their application to city systems are discussed here. According to physicists, the exponent value of power laws could differentiate the urban activities that are liable to achieve scale economies, i.e. those with exponent values smaller than one, from those that are merely proportional to the population because they meet universal needs, while others, with exponents greater than one, are seen as being accompanied by increasingly rapid growth and the risk of crises. This cross-sectional interpretation in terms of the longitudinal trajectory of an individual city assumes that the city system is ergodic. Yet this hypothesis is not consistent with an evolutionary theory of urban systems integrating the spatial distribution of labour and the hierarchical diffusion of innovation.