According to Mattern, the city-as-computer metaphor is based on the exclusionary ontology of decision trees. That grounding makes discrimination inevitable, as decision trees split issues to make an unequivocal decision rather than looking for common ground. Mattern’s proposition builds on—or, better, grafts onto—the image of actual trees. At a higher resolution, Mattern proposes the image of intermingled branching; a city should be configured through arboreal decision-making. Within the new grafting metaphor of cities as intermingled trees––in short, forests—Mattern plants two advantages. First, the grafting metaphor invokes the uncountable actors that intertwine in municipal decision-making processes. Second, with grafting, new trees begin as branches of existing trees. Hence the metaphor fosters an idea of care and interdependence between social and physical infrastructures.

Compactly written, A City Is Not a Computer fuses four articles, previously published in Places Journal, with two new chapters that synthesize all the texts into the proposition for an arboreal urbanism. The introduction extends on the title by discussing Christopher Alexander’s famous article “The City Is Not a Tree.” For Mattern, Alexander’s formal analogy between physical and social networks becomes an entry point to compare data structures and assess their embedded cultural bias. The first two chapters catalog the failures of the computer metaphor; in the first, she looks at computer interfaces, and in the second, she intellectually challenges computational metaphors. The following chapters turn to the arboreal city grafting metaphor. They open with Mattern’s most salient case study, the library, which digitization transformed from a book depository into a place for public knowledge-making. Becoming centers for people to meet and communicate, libraries turned into accelerators in which people can engage creatively in forms of bottom-up community building.
 

The recent success of libraries lies in the robustness of the physical infrastructure on the one hand and the versatility of its use on the other. Mattern claims that in the Smart City paradigm, digital devices are deployed citywide to control and optimize traffic flows. In contrast, the public initiatives that evolved in public libraries deployed digital tools ad hoc, limited to distinct periods, places, and groups of people. The spatiality of the library, meaning the mix of open plan and open access, enabled people to gather for specific projects but also to pursue other paths at some point. The library thus becomes, for Mattern, a metaphor of an open-ended city through the physical structure’s capacity of placemaking.
 
Arboreal urbanism, which is the topic of the last chapter, is a process (or discourse) that enables “branching” between physical and social infrastructures. Successful arboreal urbanism is predicated on intentionally not seeking to describe a city in its “messy totality,” but as an ongoing project of care that branches emerging social networks from urban infrastructure. Urban design, therefore, turns into the maintenance of codes to provide a robust ground for the constantly changing ways of living together in a city.

According to Mattern, deeming the city as a computer that predicts interpersonal relationships only continues ancient top-down politics that replace the prismatic complexity of cities with singular narratives. Although prismatic complexity is not credited to computers in the sense of a Turing machine, a computer is a universal machine capable of processing and storing any kind of information. Mattern’s critique is only directed against the metaphor of the computer as a performance-oriented machine. One could also ask: Why wouldn’t computers be the ideal metaphor for cities? What new metaphors would follow from a prismatic and less binary analysis?
 
The problems addressed in the book are rooted in the local context of the author; the book begins and ends with examples from New York. With a vivid study of Wi-Fi spots in New York’s subway, Mattern visualizes the smart city as a panopticon of surveillance mechanisms. She gives the Wi-Fi metaphor a positive angle: Wi-Fi can link you with distant friends, a great relief within a tunnel filled with strangers. Mattern concludes her critique on smart cities looking at the new Hudson Yards development that became possible through the misappropriation of public funds. One quality of vivid metaphors is that they can easily be flipped upside down and countered with their own prismatic realities. For instance, contrary to Mattern’s argument, corruption might be prevented through more digital transparency, i.e., data tracking and surveillance.

When Christopher Alexander saw in modernist city planning a tree, his genius move was to equate a city’s traffic arteries (another metaphor) with its web (and another metaphor) of social relationships. Yet Alexander’s metaphor reduces the plurality of interpersonal communication to the physics of a traditional, pedestrian city. According to Alexander, the mathematical model of a semilattice, visualized as a network of streets, contains multiple intersections at which people can coincidentally meet and interact. However, if people meet at an intersection while driving cars, they typically do not talk to each other. The number of intersections is not relevant in a car-based city. Although referenced in every class on urban design, Alexander’s seminal essay rarely influenced the dominantly car-based development of cities in the United States. Would Alexander’s impact on US urbanization have been different if he did not revert to a singular metaphor? What emerges when metaphors are dropped entirely, and a tree or lattice only compute data? With ease, a search engine can draw multiple facets of a city. What actually distinguishes a graph-based data structure that links any content, like Google Search’s PageRank algorithm, from arboreal notations?

Mattern rightly points out that Smart City applications that follow the traditional practice of a master plan have repeatedly failed—for example, Alphabet’s recent Toronto Waterfront development. However, Alphabet, Google’s parent company, also markets a whole range of smart-home devices. Driven via an app, such as Google Home, those tiny switches help to light, cool, and heat homes efficiently. Google Home has now been downloaded over 100 million times. Who before in history could temper 100 million homes at once? Google Nest has claimed that without a control room or master plan, cities’ carbon emissions began measurably to drop only by way of the app’s hints shown on smartphones held in people’s palms about how to save energy. Grafted onto existing urban infrastructures, didn’t Google Home branch a new urban network?
 
Mattern’s book opens a much-needed ethical debate. Planting on her argument, we have to graft a prismatic analysis of “computer cities” onto existing infrastructure to identify their opportunities and to set their limits; to find ways for a grafted care of purely algorithmic models; and to grow a city that is not a computer.
 
Daniel Koehler is an architect, urbanist, researcher, and cofounder of lab-eds. He is an assistant professor at the University of Texas at Austin School of Architecture. His current research focuses on the urban implications of distributive technologies.
 
How to cite this: Koehler, Daniel. Review of A City Is Not a Computer: Other Urban Intelligences, by Shannon Mattern, JAE Online, December 6, 2022.