Atlantic Cities

Embracing the Autocatalytic City

Embracing the Autocatalytic City

The historic diversity of the city — the source of its value and magnetism — is an unplanned creation of many hands and long historical practice. Most cities are the outcome, the vector sum, of innumerable small acts bearing no discernible overall intention.

— James C. Scott, Seeing Like a State

Cities have the capability of providing something for everybody, only because, and only when, they are created by everybody.

— Jane Jacobs, The Death and Life of Great American Cities

We are living on an urban planet; our cities are growing at spectacular rates. This growth has created new energy and excitement (cities account for 70 percent of the global economy), and it has highlighted the dysfunctions of cities. Most of our cities, particularly the fastest-growing ones, are messy, confusing places, even for the citizens who call them home. From the massive week-long traffic jams in Beijing to the crowded favelas of Rio de Janeiro, urban dwellers everywhere can easily rattle off a list of what doesn’t work in their communities. The call to action is always the same: "Better planning, better management!"

That call, though, rests on an unquestioned assumption about cities. In this modern age, we think of cities as large institutions or machines. We talk about their failures as failures of management, coordination, governance. We think we could have "better" cities if we could only tune the machine to make it more "efficient." The machine model is implicit in the popular language around "smart cities." The promise is that shiny, smart boxes will figure out how to make our cities tick by smoothing traffic flow, monitoring crime, and allocating power through smart grids. Cities will be run by supersized versions of 2001: A Space Odyssey’s HAL crunching continuous streams of big data. As Donald Fagen of Steely Dan sang, "A just machine to make big decisions / Programmed by fellows with compassion and vision."

We need to think again. In reality, these deterministic paradigms are dwarfed by the scale and complexity of our cities. Urban centers are evolving organisms, not engineering problems. Although we are able to control parts of a city — central business districts, mass-transit systems, water distribution — we will never hold and understand the whole. Cities are dynamic, complex-adaptive systems composed of millions of relatively freewilled individuals who each day make hundreds of individual decisions that set in motion consequences leading to a million other decisions. This stochastic chain of choices adds up to an emergent whole.

And these bottom-up processes — if we respect and embrace them, rather than try to control them — create cities that actually work for the people in them.

•       •       •       •       •

Picture a city whose streets are narrow and winding. The tangled network branches and turns, some turnoffs leading to even narrower lanes, many leading to dead ends. There are no sidewalks, no front yards, and nearly every door is a business — a store or a workshop. Apartments and homes sit on top of the workplaces; old and new stand cheek by jowl. The buildings crowd into each other in a cacophony of materials. A house made of concrete or brick stands next to a workshop in cinderblock; the next is wood.

There are places to eat every few steps. Corners host small markets and vendor stalls. The street itself is crowded with people. Some buy from the stalls and stores, many carry loads to be delivered from one workshop to the next. The air is thick with greetings, a friend calling out to another friend, a neighbor or co-worker nodding to an acquaintance, regular customers haggling with vendors.

Most likely, there are no maps of this place, and not all of the streets have names. In this part of the city, you do not find a location by asking, “Do you know this house number on this street?” Instead you ask, “Do you know where so-and-so lives?” or “Do you know the shop that sells such-and-such?” And someone will know. There are signs advertising services or jobs, signs upon signs — for plumbers or painters or music lessons.


Can you identify this city? Its close-knit commerce and community could appear almost anywhere, but this is the Shimokitazawa neighborhood of Tokyo. Image courtesy of Save the Shimokitazawa

Where is this city? If the signs are in Japanese, then you might be in Shimokitazawa or some other district in greater Tokyo. If they are in Hindi, you could be in Dharavi in Mumbai. If in English, you could be on Kitengela Road in Nairobi; if in Italian, perhaps in old Bellagio or the heart of Torino; if in Portuguese, possibly in Mafalala in Maputo, or in Rocinha in Rio. You could be in the old section of a European capital or in a slum on the edges of a city in South Asia or sub-Saharan Africa, or in the heart of a metropolis in Latin America. 

This city is ubiquitous and is what Matias Echanove and Rahul Srivastava of the Institute of Urbanology call the default mode of urban development. It is what human settlements that grow up without central planning and control look like: seemingly chaotic, labyrinthine, and fractal, but thick with social and business networks. Echanove and Srivastava have put together startling images of street scenes in Dharavi (the largest informal settlement in Mumbai) overlaid onto a street in Tokyo or Torino. The texture of the built environment in these disparate places is similar because the dynamics that drive it are similar. The result seems unruly, but it works. The leather workshop uses buckles made by the smithy next door, which also supplies the bag maker around the bend. The workshops are also stores, and the laborers and clerks are just a few steps from eateries. Many live in or above the shops where they work. This is the urban economy and the urban supply chain at its finest, most dynamic grain.

The default mode of urban development is autocatalytic, driven by the economic logic of proximity and supply and demand.




When images of their streets are spliced together, it can be hard to discern the difference between Mumbai’s Dharavi slum and Torino, Italy (above), or between Dharavi and Tokyo. Images: Matias Echanove and Rahul Srivastava

With the exception of water, sewer, electricity, and other services that require government, the default mode of urban development provides what is on any planner’s checklist for a sustainable and livable city: density, walkability, mixed uses. It builds engines of local development and commerce; the income generated by the shops mostly stays in the neighborhood, either as consumption of other services or investment in upgrading tools and buildings.

The autocatalytic city also confounds planners and city managers. Authorities have always struggled to control the slums. (Baron Haussmann built wide boulevards in Paris expressly to prevent barricades and allow the royal troops to move quickly across town to quell rebellions.) And so officials ignore these areas in policy and planning, or move to demolish them, often with violence. Few cities have taken stock of how much informal settlements contribute to the urban gross domestic product. By some accounts, Dharavi constitutes as much as one-quarter of Mumbai’s economy.

In contrast, the machine model of the city, when taken to the extreme, has produced aseptic cities like Brasilia and Islamabad and Chandigarh — places that are aesthetically orderly, especially when viewed from the air, but have very little life on the street. Most of these cities are ringed by autocatalytic towns where vibrant urban life actually happens.


Brasilia, the capital of Brazil, was deliberately developed by an urban planner, an architect, and a landscape designer in 1956. (Gary Yim/Shutterstock)

•       •       •       •       •

The autocatalytic city contains an intelligence, a kind of ingenuity that can never be captured by a top-down system of control. So it is almost poetic that the complexity of the city finds an analogue and an ally in the nonhierarchical complexity of the Internet. In much the same way that the autocatalytic city makes maximum use of physical materials and space, it is also co-opting technology into its fabric.

Community-based groups like Shack/Slum Dwellers International have enlisted citizens to conduct their own censuses and create their own maps, which they publish on open online platforms, thereby expanding our collective knowledge of cities. Map Kibera, for instance, used youth volunteers with handheld GPS units and the wiki OpenStreetMap to create a detailed map of the most famous informal settlement in Nairobi. Until Map Kibera, most government maps showed the area as a forest. The project revealed the network of paths and roads and showed locations of churches, clinics, and stores. Residents of Kibera are now using the map as a platform to report uncompleted or badly built government projects, countering official reports and often exposing corruption. Similarly, Transparent Chennai helps informal settlers in the city formerly known as Madras to map their own settlements in relation to (the lack of) government services. These efforts combine local knowledge with technology to engage planners and city leaders.


The desolate, official map of Kibera, shown at left from Google Maps, reflects nothing of the dense life visible in Google’s satellite view. These two images were accessed on the same day in December 2012, just minutes apart.

The use of technology is, like the autocatalytic city, built up incrementally responding directly to needs.

Nowhere is the power of this process more pronounced than in transportation. While services like Uber, Waize, Zimride, and Zipcar are disrupting the established regime in the developed world, entrepreneurs in emerging markets are also using information technology and cell phones to radically reinvent transportation, improving services for users and boosting the livelihoods of drivers.

Unlike most cities in the U.S., urban centers in the developing world are transit rich. Informal public transit permeates the urban fabric. Just as the built environment in the autocatalytic city is driven by bottom-up processes, the need to move around in rapidly growing cities with inadequate public transportation has given rise to private transit services (also called informal or paratransit). Even supposedly egalitarian public transit can be out of reach for the poorest urban dwellers. The vast majority of riders on Delhi’s much-touted new subway, for instance, have incomes more than three times higher than the local median.

Each metropolis has its own versions of the same paratransit vehicles: small buses, jitneys, three-wheelers, and motorcycles. There are the “trotros” in Lagos; a relative of the “matatus” in Nairobi; and the megataxis in Manila. The “collectivo” minibus of Latin America is similar to the Philippines’ “jeepney” and Pakistan’s minibus. The auto rickshaw is nearly ubiquitous in South Asia. Pedicabs, “trisikads,” “becaks,” “trishaws,” and cycle rickshaws have the same configurations across Southeast Asia. The “ojek” in Jakarta occupies the same service niche as the “okada” in Lagos and Freetown, the “motodup” in Phnom Penh, and the “motorsai rapjang” in Bangkok.

In many cities, as much as 80 percent of the population depends on such informal transit. But the meetings of rider and driver have been, until now, up to chance.


Chai wallahs fulfill a critical dispatching role for Fazilka Ecocabs’ innovative service. Image courtesy of Fazilka Ecocabs

Fazilka Ecocabs, named for its small hometown in the state of Punjab, India, has a clever system. It takes requests via call or text message, and relays them to the cell-phone-equipped “chai wallah” (the guy selling tea on the street) adjacent to the nearest queue of rickshaws. The chai wallah directs a cycle puller to a waiting customer. The customer receives a text message with the expected time of arrival, the fare, and the driver’s name. This makes the service safer and more predictable for the customer. In return for signing up, Fazilka Ecocabs provides the cycle pullers and their families with free visits to the doctor, educational support for their children, and traffic safety training. Ecocabs and similar services are starting to roll out apps to prepare for the influx of smartphones. They could create services that will not only dwarf the customer base of Uber or Zipcar, but also improve the trifecta of environment, economics, and equity.

•       •       •       •       •

This essay, of course, is partly polemical. Our understanding of cities has been shaped by our Industrial Age expectations of institutional control. As urban centers boom around the globe, however, we are hitting the limits of the machine model of cities. Metropolises are growing too fast for our old institutional models to work. Our task, as so ably argued by author and urban activist Jane Jacobs, is not to command the city but to understand the processes that make it work.

Rather than dreaming up ways to control the autocatalytic city, planners and city leaders should think instead of how to enable it. We must avoid confusing aesthetic order with actual order. We must recognize the native intelligence and resilience of autocatalytic communities and not suffocate them with our push for the logic of efficiency. Our plans should be additive rather than destructive and should respect and imitate the incremental (and fundamental) forces at play.

The autocatalytic city does have limits. While optimized for the small grain and the economics of proximity, it is blind to larger-scale challenges and longer-term threats, such as earthquakes or climate change. This city can build its own streets but cannot build mass transit. It can build a local economy, but not protect a local ecology. City leaders will have to find the balance. We all stand to gain, but only if we are prepared to give up control.

This essay appears in the ebook "City 2.0: The Habitat of the Future and How to Get There," co-produced in partnership by The Atlantic Cities and TED Books.

Keywords: Urban Planning, Growth

Benjamin de la Peña is associate director for urban development at the Rockefeller Foundation, where he works on issues related to cities, informal settlements, transportation, information technology, and urban science and policy. All posts »

Join the Discussion