Berghauser Pont and colleagues construct urban comparison through multivariable morphological types rather than inherited labels such as “organic”, “planned” or “grid”. Their method combines street-network centrality, building-density measures and unsupervised cluster analysis to identify recurring spatial profiles across Amsterdam, London, Stockholm and Gothenburg. The iconic move is to make typology repeatable without making it rigid: types emerge from relations among measurable variables and can therefore describe both common and exceptional configurations across scales. This provides a precise operational grammar for comparing cities while avoiding fixed morphological stereotypes. Yet the work also makes visible a broader epistemic tension: standardised descriptors gain portability by reducing the experiential and historical thickness of each place. The paper bridges urban morphology, space syntax, GIS and data-driven classification, and is especially valuable as a counterpoint to ethnographic comparison. It demonstrates how comparability can be engineered quantitatively, while leaving open the question of what such engineered equivalence cannot register.
