What makes us think it's a city?

This paper explores the way the morphological and topological properties of built environments are perceived by people. We present a survey that examines how diff erent artifi cial environments are perceived, and to what extent (if any) they resemble urban environments. For that, we developed a series of artifi cial environments, based on scripts that control the formation and distribution of diff erent att ributes on a plateau that represents an urban environment. We employed two basic street-networks: a non-orthogonal network and an orthogonal grid. In addition, we used two diff erent size distributions that obey either powerlaw distributions (with diff erent exponents) or Poisson distributions (with diff erent values of λ). Our fi ndings show that the mental images we have regarding urban environments are diff erent for diff erent representations of the environments (2D/3D). We also discovered considerable differences between the way people with or without an educational background in planning perceive and conceptualize urban environments. Lastly, we found that orthogonal street-grid and powerlaw distributions of sizes are the strongest predictors of the similarity of an environment to an urban one. The combination between them suggests that what forms the mental image of the city in our minds is not a single feature, but the combination between top-down (the orthogonal grid) and bott om-up processes (the power-law distribution).