Jing'an Park at dusk in Shanghai, China.
Image: flickr/Robert S. Donovan
In 1961 urbanist Jane Jacobs didn't pull any punches when she called city planning a pseudoscience. "Years of learning and a plethora of subtle and complicated dogma have arisen on a foundation of nonsense," she wrote in The Death and Life of Great American Cities. Fifty years later the field is still plagued by unscientific thought, according to urban theorist Stephen Marshall of University College London. In a recent paper in Urban Design International, Marshall restated Jacobs's observation that urban design theory is pseudoscientific and called for a more scientific framework for the field.
Although urban design theory is unscientific, Marshall wrote, it is not because the ideas are based on nonsense—many of the classic urban thinkers used observations and small pilot studies to describe how cities work. Jane Jacobs, for example, proposed that a city needs four ingredients to be exuberant: mixed uses, short blocks, buildings that vary in age and condition, and a dense concentration of people. "At the core of this book is a four-part hypothesis that is demanding to be tested," Marshall says. "But when I went to look to see if it had been tested, there was virtually nothing." The problem with urban design, he adds, is that its theories are untested, yet accepted as fact. Marshall proposes to overhaul the way that urban design incorporates science into its fabric, calling for more and better urban science, and for the theories to be challenged with alternative hypotheses and rigorously tested.
Some researchers are already studying cities in scientifically valid ways, but much of this work is being done by physicists and mathematicians who have little use for urban design theory.
"In urban planning, we're like physicists without a particle theory or doctors without a germ theory," says Michael Mehaffy, an urban designer at Portland, Ore.–based design company Structura Naturalis. "We don't have a unifying idea about the nature of what we're looking on. We say we're artists but it's as if we're medieval doctors with potions…. We need to recognize that we have a responsibility to use models that are more likely to produce better outcomes."
Urban designers traditionally have doubted the role that science can play in describing or predicting or fixing a city. They assert that cities have an emergent complexity that results from the interactions among people as well as between people and the environment, and that there is an element of human behavior that cannot be reduced to an equation. "I'm not really sure that the city can be studied and predicted any better than we can predict the weather," says Peter Laurence, an architect at Clemson University. "And cities are more complex than the weather."
Geoffrey West, a physicist at the Santa Fe Institute who studies urban growth, agrees. People who study cities "are dealing with maybe one of the most complex systems in the universe," he says. "The idea that you could reduce it to an equation is extremely hubristic." Yet he and others are using computer simulations, modeling and mathematics to begin to tease out the simple rules that give rise to urban complexity and diversity. For example, West has devised formulas that use a city's population to predict the size of the metro police force, the number of people with AIDs and the annual tonnage of carbon dioxide emissions. The predictions are accurate 85 percent of the time. That consistency is guided by universal social dynamics that govern the way people interact with each other and with the environment, West says. The other 15 percent is individual variation between cities.
Rights & Permissions
See what we're tweeting about
18 Comments
Add CommentIt is always annoying to me when people claim their particular area of interest is outside the realm of science. Balderdash.
Reply | Report Abuse | Link to thisIf you are trying to solve a problem, what's better, to apply a proven problem solving technique, or to just guess/go-with-your-gut? The intuition tactic might hold its own at the beginning while the scientific approach gets organized, but over time, science will beat every other method hands down.
Sure the problem may be too complex to be fully described by current science, but science offers the best shot we have at getting the closest to the correct answer, and the thing about science is, as long as you continue to work at it, it gets closer and closer to the optimal answers over time.
If you apply the scientific method to city planning, cities in the next century will be dramatically better for it. Will it be perfect on day one? Clearly not, but a journey of a thousand miles begins with the first step.
While that annoys me as well, I do not believe that is the case here. There are a number of issues that make it difficult and one misnomer is they are practicing science, however many of the modern methodologies and considerations of modern science are lacking.
Reply | Report Abuse | Link to thisThe difficulty is in the codification of the knowledge, which either isn't set to how modern science is practiced or the relationship between theory, observation and experimentation are all confounded.
The use of computational simulation tools is certainly a positive step but the scope of urban design hurts further progress because of the number of experimental parameters and variables that are closely related.
As these are designs and relate to mass and individual human psychology efforts are further hindered. Make declarations about the objective superiority of one design is very difficult because you have to contrast it with the subjective performance of countless other iterations that happen in objectively different environments over huge longitidunal spans with missing and incorrect data.
I'm not saying there is no hope, nor that science is of no use. Just that it will take time to paradigm to shift within the urban design community and for more wide spread use of computational tools.
I await with great anticipation for a science based evaluation of the following important questions:
Reply | Report Abuse | Link to thisIs Callas superior to Sutherland?
Are the Stones superior to the Beatles?
Is Van Gogh superior to Rembrant?
Is atheism superior to religion?
Once these basic questions are resolved by science, perhaps science can move on to determine the best urban form.
I think you're wrong; systems analysis for instance is certainly an applicable discipline because cities are collections of interacting systems that constitute a larger system. All cities have many subsystems in common; industrial areas, residential areas, "office spaces" both private and public, retail and wholesale marketplaces of various sizes and types, transportation networks that move people and materials from one place to another within (as well as into and out of) the city, and other infrastructures that support all the activities of a city such as utilities and government facilities including fire and police.
Reply | Report Abuse | Link to thisAs I see it the problem is not the so-called intangible "human factors", it is an incomplete understanding of how all these subsystems can interact optimally, local geographical, environmental, religious and political peculiarities notwithstanding.
It might be useful for urban planners to first think about defining their terms- what is a "city"? Which components listed above (and which ones not listed above) are critical to a city's function?
It might also be useful to consider actually following the scientific method explicitly by first making observations of established cities to see what works well and what doesn't, and then try to discern *why* those that work well, work as well as they do *before* theorizing how to "design" a city.
Artistic sensibilities notwithstanding, form must follow function. What is a city supposed to *do*? Answer that question first, urban designers.
A basic idea that provides the foundation of science is that there is (are) single knowable truths. We assume that the physical world operates under a single set of rules and the we can describe the universe. In speaking of 'solving a problem', tharriss invokes this idea. For most all of our physical sciences problem solving to uncover some single truth is fairly valid.
Reply | Report Abuse | Link to thisThe complexity of a City cannot be boiled down to one set (or variations) of rules. Cities exist within the interplay of all the diverse perceptions of the occupants. The direction that would be 'good planning' for Duluth would most likely make no sense at all for Kathmandu.
Many city planners (that I have worked with over the years) like to think that planning is scientific, but the reality is that they are always trying to plug round pegs into square holes. Over the years I see planners applying the new urban concept du Jour, which is sweeping the country. What they should be looking out is how the dynamics of their individual situations function.
It's not clear if they are trying to draw a hard line between urban design and urban planning. There is plenty of scientifically valid research in the urban planning field some of which overlaps into urban design issues. At a very basic level, WalkScore is demonstrating a relationship between their index and property values that is being noticed in the private sector. Reid Ewing's work is worth looking at. There's also Space Syntax. I recall seeing research on how long people idle in different urban design contexts.
Reply | Report Abuse | Link to thisA scientifically valid study doesn't have to involved modeling. Modeling brings in a whole different set of challenges.
There is however a legitimate case to made that if the statistics seem to contradict everything you've encountered on an anecdotal basis, the statistics may be missing a factor or there may be a factor that is difficult to quantify.
The scientific method shouldn't completely overshadow common sense or simple observation. It's akin to trying to program a computer to recognize an image versus asking a person to do it. It's not impossible to program a computer to generalize an image but a person can do it really quickly often at lower cost.
@metahominid Isn't a basic element of the scientific method having a hypothesis to test? At least that's what they taught us in grade school.
Reply | Report Abuse | Link to thisThere is a fundamental difference between social "sciences" and natural sciences! Try as we may, human behaviour cannot be controlled, so there is no chance that social theories will ever be "verifiable" in the scientific sense. Just because one applies the tools and techniques of science to something does not make it "scientific".
Reply | Report Abuse | Link to thisWe need to be making more effort to understand what models and plans can achieve, and what they cannot. Arguing about whether or not urban planning is "scientific" is counter-productive in my opinion.
The scientific method is basically a set of steps to find answers to a question; observe, hypothesize, predict, test (experiment), and analyze.
Reply | Report Abuse | Link to thisWe can't legitimately form hypotheses without adequate data, and it seems to me that people who want to design "better cities" should first figure out what the scale quantifies and how to measure it. As I said earlier, it might help to look at existing cities and let the cities tell designers/planners what criteria the inhabitants of those cities considered important, rather than trying to impose fanciful abstract design rules on people's lives.
I'm not sure why believe I'm wrong. Perhaps I should have been more explicit with each thought and the conclusion in particular. I am on your side, you should note that my whole reply to tharriss is in agreement but that cautionary.
Reply | Report Abuse | Link to thisI was attempting to make him pay attention Geoffrey West's mention. These human factors are not intangible, they are frequently unseen but the most important factor. The people in the city make the city more than any skyscraper or bridge. Those are simply ways shaping their interactions.
I am all for the use of scientific method but like are you pointed out (and the article) that without a more direct and singular understanding of the city, or a coherent unifying scientific underpinning, then much research progress become nebulous.
I get the feeling that, your condescencion of their research methodology aside, you are forgetting that 'better' is ill-defined. The work which is scientific has often expressed using a few scalar variables and parameters with end up being more or less two dimensional engineering characteristics of fiscal or population growth. Getting long-term high dimensional data to actually attempt to reproduce, something modern science is accountable for, the efforts is obfuscated by so much red tape.
We are not disagreeing, I am just saying from experience this a very difficult thing to do.
@myurbangen - I'm sure if the part about requiring modeling was directed at me or in general. I understand that, but modeling is especially useful in this context due to the difficulty in reproducibility from economics and free will (people don't just build their city how you state in a research paper). Experimental design needs reproducbility modeling is cheaper than the alternative and offers the ability to hypothesize about events which cannot or will not occur.
Reply | Report Abuse | Link to thisIn response to the snarky grade school part, perhaps it was implicit but when you model you actually must direct the constraints of your model to align with a pool of possible hypotheses from the experimental construct. There are no programs, that I know of for urban design, which autonomously gather and synthesize observational data to generate notable models for the testing of hypothesizes. Each researcher must actually program the model to test what they are atttempting undestand.
I don't know what planet Marshall is living on. Numerous studies have been done that demonstrate that denser, walkable, mixed-use communities have residents that are happier, healthier, and use less energy than their suburban counterparts. Sounds like poor research on Marshall's part.
Reply | Report Abuse | Link to thisI wasn't trying to be condescending, and I pointed out that design and planning are both pointless unless we have some idea what we're designing and planning for. That's why I raised the questions of what a city is and what it's supposed to do; why do we build and live in cities and what do we want from them?
Reply | Report Abuse | Link to thisWe (as a species) have built and lived in cities for a very long time, with and without deliberate planning. Some grow organically around a certain set of activities and others get tacked on, some successfully, others not so much.
Cities are not as you say bridges and skyscrapers any more than they are collections of mud huts or imposing piles of stone; they are places for people to interact in specific ways. In that sense I disagree with yarberry; there are constants across all sorts of cities. In my opinion we must focus on the interactions and how to facilitate them.
Again though, to hypothesize, first data must be collected. We should study examples of cities, both ancient and modern, to see what works and what to avoid. To know where we should go, we must know where we've been and where we are.
Ok, it seems like were fairly on key here then.
Reply | Report Abuse | Link to thisI don't think your intent to is convey that there is little data being collected because in all of history there has always been different forms of data collected by the local officials. There are certainly purely academic pursuits. I think what will be useful and likely in the next few years is the synthesis of tools for academics and officials on a wider spectrum to communicate and also communication with corporations to facilitate Big Data.
The problem is getting private citizens to help make the process easier.
"We don't have a unifying idea about the nature of what we're looking on." maybe because there is no unified object called "the city" ;-) this is not a new idea... there's a reason all the attempts to create scientific laws of 'the city' in the past have failed miserably. read Richard Sennet's op-ed in the guardian on "the stupifying smart city'
Reply | Report Abuse | Link to thisno one on this thread understands, scientific urban planning has been around for a long time, perhaps before the Chicago school. Everyone who has tried has either distilled the 'city' right out of their model, or worse instigated some horrible e smaster-planned 'science cities' that have destroyed large swaths of the urban environment. yesterday it was garden cities, today it is smart cities but the same mistaken universalising logic favoring narrowminded goals of 'efficiency' or somesuch still applies today. I hope I never have to live in a place designed by 'scientists'
Reply | Report Abuse | Link to thisThis is an interesting discussion, to which I'd add the comment that city planning and urban design are different fields and disciplines. (This is somewhat confused in the headline of this article.) Nevertheless, it seems difficult to deny that both should be grounded on substantiated thinking. However, it is also the case that regional planning and city planning tend to be practiced less as design fields than data-driven social sciences. By contrast, urban design is a design field, as the name indicates, and is more of "an art and a science."
Reply | Report Abuse | Link to thisStephen Marshall's fundamental and I think most important point with regard to urban design—or, more specifically, urban design theory, as he distinguished it from urban design at large in his article—is that urban design has been treated too much as an art and has lost (or never had, if you follow his logic and argument) a "scientifically" verifiable basis.
Again, it is hard to disagree that urban design or urban design theory should not attend to the science-side of its "art and science" whole. After all, medicine is also an art and a science that relies heavily on science, but also the art of the doctor's interaction with the complex things that they deal with.
And this was exactly Jane Jacobs's point with her comment about city planning as pseudoscience and her comparison of it to the life sciences. It was also why she raised the issue of complexity science in her thesis about the complexity of cities—and in doing so was one of the first people outside of the sciences to make the connection between complexity science and any field outside of mathematics, decades before the development of that field. (This, incidentally, makes Marshall's critique or implied critic of her rather contradictory.)
For Marshall's article see http://www.palgrave-journals.com/udi/journal/v17/n4/full/udi201222a.html.
For more on urban design history as related to Marshall's article, see http://www.tandfonline.com/doi/abs/10.1080/13574800600644001.
For more on Jane Jacobs and complexity, see http://onlinelibrary.wiley.com/doi/10.1111/j.1531-314X.2006.00033.x/abstract.
I agree with Zvi.Leve. Interesting article but it does not go in depth of the whole problem, (usually that does not happen in Scientific American), and that is with the whole apparatus the Social Sciences work with, their problem of methods, tools & materials. As Urban Planning & Design is/are an applied social sciences we have to get into that discussion. One one side we need a completely new way of working with urban design, in terms of methodology, sort of a new grounded theory for the profession and also a new theory of urban design that will accompany this. I know how this sounds but my students of urban planning & design have been saying to me this for 6 years - 6 different generations, and I agree fully with them. All Kudos to Space Syntax & beyond, theory of complexity, systems analysis, and all the other stuff around including, Jacobs/Whyte/Gehl/Lynch/Cullen/Sitte and dozens of others that you all knew...the professions are stuck and they cannot respond to the complexities of the urban development and the converging crises we have today. We have an enormous task ahead of us and calling our disciplines pseudoscience or no science at all will not get us anywhere...
Reply | Report Abuse | Link to this