Currently
   07 Emergence

Past issues
    06 Memory
    05 Revive
    04 Hidden City
    03 Anthropocene
    02 Fake
    01 Resistance

Info
    About
    Contact
Mark




CITIES, COMPLEXITY

AND EMERGENCE:
WHAT NEXT?

April 2025







by Richard Shearmur

Complex systems
There is a fascinating body of knowledge, emerging since the 1970s, on complex systems. Complex systems can be approached through the mathematics of fractals and agent-based models, but first and foremost they are a way of thinking about the world. Just as the Enlightenment (precursor to the industrial revolution and the modern age) viewed the world as a mechanism – governed by knowable and immutable laws of physics and biology, which science would uncover – so complex systems theory invites us to see the world as an anthill, made up of  interrelated objects (the material world) and agents (people, animals,…ants), governed by laws of physics, society and biology that are not fully knowable, and that, even if knowable, would not enable prediction or control.

Just as anthills emerge from the uncoordinated actions of ants as they interact with themselves and the material world following simple behavioural patterns, so cities emerge from uncoordinated (or partially coordinated) actions of people as they interact with themselves and the material world following behavioural patterns that do not, in any straightforward way, lead to the cities that we see around us.

Ok, but what does this mean in plain English? A ‘system’ is an ensemble of objects and agents that interact. The mechanical view of systems is that, once we know the rules of interaction and the initial situation of each object and agent, then the system is predictable and understandable. The complex (or anthill) view of systems comprises the following elements:

  1. We can approximately know the rules of interaction that govern a system.
  2. We can approximately know the initial situation of each object and agent.
  3. These approximations are irreducible (there will ALWAYS be some level of approximation, if only because of Heisenberg’s uncertainty principle, mentioned below).
  4. Because of these approximations we have limited capacity to control and predict.
  5. Some complex systems are stable, leading to outcomes that, though unknowable, fall within certain parameters; some complex systems are unstable – they are chaotic and unpredictable (even though we may understand their basic rules).

Cities are complex systems
Now, those of you who are still with me may be saying to yourselves: “Surely, a city is a system. Conceptually it consists of objects and agents that interact.” You are correct.

And since you are correct, the next step is to realise what this entails (if we accept that systems are complex and not mechanical): cities – and urban processes – are inherently unpredictable. Our incapacity to envisage ‘unforeseen consequences’ is not solely due to incompetence: it is also, partly, due to inherent unpredictability.

Why can’t all urban planning mistakes and cost overruns be assigned to complexity? Because complexity theory recognizes that many systems are predictable in the short term (the ‘short-term’ is very elastic, but, for the sake of argument, a few years), provided that the system has been well-enough specified and that a meteor does not fall out of the sky. However, beyond this short-term, inaccuracies mean that outcomes become less and less foreseeable.

This view of cities places some fundamental constraints on what can be expected from urban planning, whilst opening up possibilities if complexity is embraced and planning becomes more interactive and reactive, focussing on working with communities to establish goals and visions, with less emphasis on plans and regulations.





Why don’t better knowledge and measurement get us back to mechanical cities that can be planned and regulated? 

Inaccuracies that make systems unknowable are of two types. First, the initial situation of all objects and agents in a city is impossible to know. Even if we invented a measurement technique, the time it would take to gather and marshall information means that the situation would alter during the measurement process.Heisenberg’s uncertainty principle in physics, which states that one cannot know both the speed and position of a particle (because observing its location perturbs its speed, and measuring its speed alters its location), applies to cities too.

Second, the rules of interaction are also approximations, and can change over time: this is easy enough to understand within cities, as politics, social change, technology, etc. alter the nature of interactions.

Finally, these unknowns compound each other: there are feedback loops, as agents and objects alter in reaction to how the system evolves, itself a result of (approximately known) initial conditions and (approximately known) rules of interaction.

No amount of data, AI or modeling can overcome these irreducible uncertainties.

Emergence

Anthills consist of a huge number of agents (ants) and objects (sand, food, stones...). The agents follow fairly simple behavioural rules, from which it would be difficult to predict – had we not observed it many times – the emergence of an anthill: there is no central planning nor any blueprint. Should the anthill be destroyed, a new and somewhat different anthill will emerge. Ants are an example of a stable complex system: no two anthills are identical, yet, given the basic rules and building blocks, all ant colonies will build anthills that have similar general properties.

Complex systems can operate like ants: rules of interaction can be such that a stable pattern with known general properties emerges.

It is possible, though, that an anthill will not emerge: this can happen if there are no appropriate materials (objects), but can also happen if ant behaviour is sufficiently perturbed that it ceases to produce anthills. Unstable complex systems are chaotic: their behaviour cannot be predicted, even in general terms, even if one has a good grasp of the rules that govern them. Some rules just do not lead to emergence: the three-body problem is a famous example.

What will emerge over the next few years and decades?

Since the 1950s there has been a tendency, in the Global North, to assume that society – and its cities – are stable and predictable. Rules of interaction were by-and-large understood, and were hemmed in and stabilized by institutions, laws (and planning regulations). Even big changes, such as neo-liberalism, did not – apparently – upset overall stability. Likewise it was assumed that our climate was stable and predictable: after all, environmental ‘limits to growth’ did not halt growth!

Within this context, plans could be drawn-up for 10 or 20 years, and regulations could fashion city development. With climate now changing faster than most climatologists imagined, and with institutions and social norms dissolving, our collective anthill has been kicked, and the rules of interaction are rapidly changing. Something will emerge (I suspect). But I have no idea what… maybe a first step would be to work out what we (as a society) actually want.

References:

Batty, M., 2017, The New Science of Cities, MIT Press

Beinhocker, E., 2002, Strategy at the Edge of Chaos, in Faulkner. D (ed) Critical Perspectives n Business and Management, Routledge, 32-42

Meadows, D., Randers, J. and Madows, D., 2004, Limits to Growth: The 30-Year Update, Chelsea Green Publishing

Prigogine, I. and I.Stengers, 1984, Order out of Chaos, Bantam

White, R., Engelen, E. and Uljee, I., 2015, Modeling Cities and Regions as Complex Systems: From Theory to Planning Applications, MIT Press