Got shiny new tools, but still got urban blues

This is an intriguing book. Many of the world’s leading thinkers and authors have tried to understand cities, so it was with high hopes that this reviewer embarked on an expedition into cellular automata (CA) and fractals. The expedition was not successful, and this was not because the book is a technical treatise with little consideration given to those more interested in cities than in fractals. The technicality is sometimes difficult but is not impenetrable; it is also desirable that technical matters are given this amount of detailed exposition. What is worrying, however, is that it is very difficult to find anything in Michael Batty’s book that helps us to understand cities more than we already do. This lack of progress with understanding is partly explained by a key phrase early on: "in the next chapter we will concentrate on the question of building CA models that can be applied directly to significant problems of urban growth. This will take us some way down the path to empirical applications but we will still remain somewhat aloof from the real world."

Understanding cities is about as "real world" as it is possible to get, and it is not at all clear how we can begin to appreciate cellular automata and fractals unless we can see how the theory, the models and the applications have helped to increase our understanding of cities. If it is not possible to demonstrate some improvement in understanding cities then at the very least this "undersells" "cellular automata, agent-based models and fractals".

The lack of demonstrable progress is even more perplexing when we read on the final page: "Using these ideas, there is a real sense that a better understanding can be achieved of the way our cities are structured, the way they evolve, and the extent that we can manage and design them." This suggests that there is an intention to understand and then improve, but it sits uneasily with the poor evidence base that understanding is actually happening.

There is a serious tension here. The previous 519 pages have explained little about city dynamics, spatial growth and development, the driving forces that shape cities, the circumstances that produce successful cities and those that produce unsuccessful variants. On page 152, we are told that "explaining why certain locations rather than others attract such growth is problematic" and "at a different scale the same processes that favour the growth of some cities in a wider landscape or system of cities are equally baffling". The tension lies in the stark contrast between the failure to explain and understand and the belief that a "better understanding can be achieved". There is a discussion of six US cities at the end of the book that looks at development patterns and spatial structures. The introspective discussion, entirely related to the modelling approach, leaves the reader wondering what, if anything, it tells us about Albany, Columbus, Buffalo, Pittsburgh, Cleveland and Syracuse.

Another potentially useful analysis is presented of walking in Wolverhampton. This is described as "our first foray into the real world". The foray is very disappointing, indeed. At the end of a technically impressive simulation of walking behaviour in central Wolverhampton, we have learnt nothing about walking, nothing about the links between walking and urban form and structure and nothing about what we could do with the simulation results. Equally, we have learnt nothing about how we can "manage and design" Wolverhampton or, more generally, manage and design cities based on an understanding of what pedestrians do. The whole discussion is context-free and unrelated to any policy or urban design question.

All this is in stark contrast to some of the splendid work done on cities over the past 50 years or so. It is interesting that Batty mentions Jane Jacobs and her dramatic insights into cities in The Death and Life of Great American Cities but fails to mention the even more powerful work of Lewis Mumford. He wrote profusely about the meaning and symbolism of cities, their growth dynamics and the relationship between people and cities. He then linked this to what we would now call "governance" and to a damning critique of urban sprawl. Both Jacobs and Mumford understood the complexity of cities.

It would have been so much better if Cities and Complexity at least started with an audit of what we do and do not understand and then led us through the technical landscape of cellular automata and fractals before making it clear what insights are added by these sophisticated tools. Instead, the book is a technical tour de force that takes the reader through more than 500 pages of exposition around "cells and cities", "desktop simulators", "agent and cells", "agent-based models of street systems and buildings", "the dynamics of small-scale spatial events" and "modelling urban growth as a spatial epidemic".

Mumford, in The Culture of Cities , was not afraid to grasp the nettle of design and management: "the task of city design involves the vaster task of rebuilding our civilisation. We must alter the parasitic and predatory modes of life that now play so large a part and we must create region by region, continent by continent, an effective symbiosis, or co-operative living together… on the basis of more essential human values than the will-to-power and the will-to-profits." This task remains to be done, and all of us involved in any aspect of cities and complexity need to reflect on how our approaches could contribute to the "vaster task" set by Mumford.

John Whitelegg is professor of sustainable development, Stockholm Environment Institute, York University.