Generative methods in urban design: a progress assessment

Author: Michael W. Mehaffy ^a

Affiliation:

a Environmental Structure Research Group, Lake Oswego, OR, USA

DOI: 10.1080/17549170801903678

Publication Frequency: 3 issues per year

Published in: Journal of Urbanism: International Research on Placemaking and Urban Sustainability, Volume 1, Issue 1 March 2008 , pages 57 - 75

Abstract

Introduction

Just recently we passed the 20-year anniversary of the publication of a slim and influential volume titled A New Theory of Urban Design (1987). In it the mathematician, architect, and theoretical iconoclast Christopher Alexander sought to establish “a new theory of urban design which attempts to recapture the process by which cities develop organically” (Alexander et al. 1997, p. 2).

This organic development, writes Alexander and co-authors, “is not a vague feeling of relationship with biological forms. It is not an analogy.” It is, they say, a specific structural quality: “namely, each of these towns grew as a whole, under its own laws of wholeness” (Alexander et al., p. 1). Alexander and co-authors then proceed to develop these “laws of wholeness” with detailed structural logic, and to propose a method by which this quality can be attained again in a contemporary context - not through a conventional kind of master plan, but through a process involving the sequential collaboration of a series of participants.

We can describe such a method as generative. That is, we cannot know in advance the nature of the geometric results that will emerge from the complex process, though we may know the general aims of the participants. We will generally avoid simplifying mechanisms such as large-scale diagrammic concepts, rigid typologies, or so-called “design partis” (i.e. schemata),1 especially in the early stages. Instead, the collaborating participants will together generate an evolving form that grows out of a complex transformation of the existing place and its people, together with all its environmental, social, and cultural factors. Such a generative process is continuous, and cannot be frozen in a standardized master plan.2

In this sense, A New Theory of Urban Design was a challenge to the very idea of urban design as an act of conventional schematic “master planning”, and an assertion that design must be a continuous evolutionary response to a complex environment of urban conditions. While the performance criteria may be clear at the outset, the designer's role is not to specify the final form, but rather the intermediate process that will generate that form.

Twenty years later, we can take stock of the developments since. Has anything like Alexander's “generativity” been incorporated into conventional urban design methodologies? What of the efforts to engage citizens in participatory design processes? What of the efforts of New Urbanists to conduct collaborative “charrettes” and apply “many hands” in the production of the result? What of other efforts to incorporate complexity into design strategies, or, in some cases, partially to abdicate the authority of the designer in the face of urban complexity (as was famously argued, for example, in Rem Koolhaas's (1995) essay “Whatever happened to urbanism?”). What have been the lessons learned from these experiences, and the opportunities still to be developed?

We will argue that progress has indeed been made, but important opportunities remain to be developed. We will point to flaws in previous efforts, including Alexander's own methods and assumptions (flaws that in many cases Alexander himself has already noted); and we will discuss more recent efforts, their promises and drawbacks, and further opportunities remaining.

Generative methods before New Theory

The holism described by Alexander and colleagues was not, of course, a new theoretical recognition of Alexander's team: indeed, the subject has been an active one in many scientific fields over the last half century or so, from quantum mechanics to cognition to neuroscience to embryology to genetics. Alexander, the Cambridge-trained physicist, mathematician, and polymath, was well aware of these parallel trends.

In fact, Alexander was one of the early pioneers of the so-called “design methods” movement, an effort to establish systematic evolutionary methods in design, through which a designer (or group of designers) can move methodically through the steps in a design process, and not merely rely upon more intuitive and unpredictable approaches, nor on rigid typological prescriptions. The field was driven by the need to cope with increasing complexity in technological problems in the era after World War II, fueled by the development of cybernetics and computer software (Broadbent 1979). It was believed that the old intuitive methods of design experts were simply not reliable enough to manage this complex challenge. Within this field, Alexander's work focused on the “decomposition” of a design problem, and the synthesis of the parts into a new form on the basis of various design inputs.

Problems with the so-called “first generation” of design methods (including criticisms by Alexander himself and others) led to the development of a “second generation”. For investigators like Horst Rittel, we would need to distinguish between “tame problems” (which were dealt with well enough in the first generation) and the more common “wicked problems,” which include multifaceted problems like those that are typical at a larger urban scale. To cope with this complexity, the new methodology must therefore incorporate an “argumentative process” within a network of issues. Moreover, the designer is no longer a solitary “expert” but a collaborator with the client and with other experts (Fowles 1977).

Yet a “third generation” of design method, around 1980, accelerated the emphasis on collaborative process, and offered a model in which the designer (or collaboration of designers) would propose “refutable hypotheses” for consideration by the client. Moreover, there need not be a single proposal for consideration; indeed, a “plurality of views” was necessary (Dulgeroglu-Vuksel 1999).

Alexander himself grew disinterested with the movement he helped pioneer, and ceased to interact significantly with its proponents. But his work clearly evolved in a similar direction, as he began to incorporate users in the design process, and as his method featured a cycle of evaluation for testable design propositions.

Alexander's development of pattern languages was also an outgrowth of his strategy to deal with what Simon (1962) had earlier termed “the architecture of complexity.” His “patterns” were in essence recurring clusters of design configuration, which could be recombined adaptively to create new design solutions. Alexander had long noted as well that the structure of cities included overlap and structural plurality. His “pattern languages” offered a grammatical system whereby the patterns could be recombined in redundant, overlapping ways.

As noted above, all these insights had their corollaries in the scientific developments of that time, the understanding of the ways that complex systems exhibited interrelations of their elements - and the implications for designers.

Generative methods and "organized complexity"

A classic exposition of the implications of the phenomenon of “organized complexity” for city planners and urban designers was made by Jane Jacobs in the last chapter of her landmark book The Death and Life of Great American Cities (1961). Its influence on the planning discipline in general, and on subsequent thinking about process and generativity in particular, is difficult to overstate. (For example, Google Scholar lists over 2000 citations.)

In a remarkably prescient chapter called “The kind of problem a city is,” Jacobs lucidly analyzed the implications of the scientific advancements that were then occurring - in particular, the understanding of complex systems in which a number of factors were “interrelated into an organic whole” (p. 432). This was important for urbanists because they needed to be sure they were thinking about the right kind of problem, and using the right tools to solve it.

Science, Jacobs wrote, had grown in the last 400 years from Newtonian, “two-variable” science, to the other extreme of statistical phenomena, in which myriad variables interacted. This trend reached its zenith in the early 20th century, and powered the phenomenal growth of industrial technology in its early phases.

Gradually, however, 20th-century science began to understand the realm in between, in which more than two variables interacted in important ways. This new domain opened up “immense and brilliant progress” for the life sciences, which has fueled much of the revolutionary work in genetics and other fields. Indeed, since 1961, progress in the field has only accelerated and delivered astonishing new revelations in embryology, morphogenesis, cognition, and many other fields outside the life sciences too. Jacobs was one of the first to spot the implications for urban design and planning.

But planning and architecture, she noted, were lagging behind - still trying to treat the problems of the city as either simple two-variable problems of simplicity (this many jobs over here, that many houses over there) or problems of statistical populations to be managed, almost as files in a drawer. They misunderstood “the kind of problem a city is” - with devastating results.

This amounted to a devastating critique on the top-down master-planning approach - from the highly influential Garden Cities movement, which sought to isolate and neatly segregate planning variables such as housing and jobs (Figure 1), to the statistically informed “towers in the park” of Le Corbusier and other modernists, relying upon statistical management of family sizes and income groups into schemes that conceptually resemble nothing so much as giant filing cabinets (Figure 2):

With these techniques, it was possible not only to conceive of people, their incomes, their spending money and their housing as fundamentally problems in disorganized complexity, susceptible to conversion into problems of simplicity once ranges and averages were worked out, but also to conceive of dry traffic, industry, parks, and even cultural facilities as components of disorganized complexity, convertible into problems of simplicity. (Jacobs 1961)

But this is a disastrous strategy, she argues, likely to miss critical organic relationships, and likely to result in dysfunctional, sterile and oppressive environments:

As long as [we] cling to the unexamined assumptions that [we] are dealing with a problem in the physical sciences, city planning cannot possibly progress. Of course it stagnates. It lacks the first requisite for a body of practical and progressing thought: recognition of the kind of problem at issue. Lacking this, it has found the shortest distance to a dead end. (Jacobs 1961, p. 439)

Processes, Jacobs says, are of the essence in cities - and in the ways we must interact with them. The neat segregation of earlier planning methods must be discarded, in place of a more diverse, more mixed model - managed not by simplistic top-down schemes, but by a kind of diagnostic approach, seeking to understand and treat the existing system, transforming it to a healthier state - almost as a medical professional would do.

In the years since Jacobs's landmark work, many authors and practitioners paid homage to her insights. Yet Jacobs reportedly found this vexing.3 Planners, in particular, were the subject of her harsh and continuing criticism for failing to respond to her critiques - worse, pretending to honor her ideas while in fact continuing with disastrous conventional policies. For her, apparently, the lessons of process and generativity embodied in organized complexity were not being taken nearly seriously enough.

In the intervening years, a number of other prominent planners have taken up the challenge to deal more effectively with “organized complexity.” Bill Hillier, Professor of Urban and Architectural Morphology at University College London's Bartlett School of Planning, developed his theory of so-called “space syntax,” a methodology that maps the connective relationships within a spatial system, and expresses the global connective properties of each element. In so doing Hillier has devised an elegant tool to show, in discrete units, the results of a complex set of connections of a number of globally interacting spatial variables. A street segment that is well connected within its global context shows high connectivity, and thus might be especially appropriate for a retail function, or a civic element. In this way, the local functions emerge from the generative properties of the global ones - exactly, Hillier argues, the way they did in many traditional organic cities (Hillier 1989, 1999).

Mike Batty, Hillier's colleague at UCL, has also developed a number of methodologies to analyze and manipulate environmental complexity. His “Center for Advanced Spatial Analysis” has used generative algorithms to analyze various organic patterns and their generative rules. Batty and his team have paid special attention to the properties of “self-organization,” the tendency of systems to develop patterns of organized complexity spontaneously as a result of algorithmic sequences of activity. Just as bird flocks form large-scale coherent patterns from simple rules of distance followed by each bird, so residents of an informal city can build roads and other remarkably coherent structures by following relatively simple rules (Batty 1991, Batty et al. 1997).

These insights parallel, and clearly draw from, the rapid developments in complexity science in general, and in particular the phenomenon known as “emergence.” Investigators have been able to identify with mathematical precision the processes that give rise to complex structures from an apparently simple set of rules, with useful implications for game theory, economics, biology, physics, meteorology, and many other fields. In the fields of planning and urban design, the insights can be used to understand the relationship between complex urban form and relatively simple generative rules - like those followed by a group of actors in a building process.

The significance of Hillier's and Batty's tools is that they have applied these insights to the urban toolkit available to conventional practitioners, in effect to re-grow the urban complexity that previously existed or could potentially exist, with desirable results. Hillier's analytical technique in particular has been put to the test a number of times with notable success. His analysis of the proposed modification of London's Trafalgar Square predicted an increase of pedestrian activity of some 16-fold. The actual results of the work, Hillier reports, were very close to that prediction.4

The publication of A New Theory of Urban Design (1987)

Twenty-six years after Jacobs's The Death and Life (1961), Alexander and colleagues published A New Theory of Urban Design (1987) - containing more than a strong echo of Jacobs's emphasis on process, and on the understanding and treatment of existing organic wholes. To Jacobs's insights, Alexander made several significant additions. Chiefly, he wanted to propose a methodology by which such a collaborative process could produce geometries that had the characteristic of organized complexity - the characteristic he identified as “wholeness.”

A New Theory of Urban Design is not among Alexander's best-known works. It is surely eclipsed by such familiar works as his landmark PhD thesis, which became his first book, Notes on the Synthesis of Form (1964); his elegant little paper on differential urban geometry, “A city is not a tree” (1965); his perennial bestseller A Pattern Language (Alexander et al. 1977), reportedly the best-selling treatise on architecture of all time; and its companion volume, The Timeless Way of Building (1979). According to Harvard Design Review Editor William Saunders, A Pattern Language “could very well be the most read architectural treatise of all time” (Saunders 2002). By comparison, A New Theory was a modest seller, and did not seem to have anything like the professional impact of the others.

Yet two decades later, the little volume seems to carry an influence far beyond its initial sales would suggest. We hear its ideas still discussed today,5 and we may find it sitting prominently on rather unlikely bookshelves. Moreover, as argued herein, it serves as a kind of milestone in the discourse on process and generativity in urban design.

Even Alexander's own assessment would not suggest that A New Theory should have the legacy it has. Alexander himself argued later that the book's focus was still too heavily upon the formal design product, and did not deal sufficiently with processes of social interaction, site assessment, financial arrangements, or construction sequencing and management. Even the basic problem of geometric form was not dealt with sufficiently, he argued. All these topics would be taken up in far greater detail in his much more recent magnum opus, The Nature of Order (2003-04).

Yet in this volume, Alexander and colleagues did squarely reintroduce the notion of process into the debate about urban design - a process that aims, above all, to generate wholeness within the urban structure:

We have therefore used the phrase urban design in the title of this book, since it seems to us that urban design, of all existing disciplines, is the one which comes closest to accepting responsibility for the city's wholeness. (Alexander et al. 2003-04, p. 3)

But Alexander wants to challenge the very notion of design, and its sovereignty over city-making as a technical exercise or a creation of “master plans:”

But we propose a discipline of urban design which is different, entirely, from the one known today. We believe that the task of creating wholeness in the city can only be dealt with as a process. It cannot be solved by design alone, but only when the process by which the city gets its form is fundamentally changed. (Alexander et al. 2003-04, p. 3)

The echoes of Jacobs are readily recognizable. But there are also echoes of Alexander's own earlier work on the overlapping networks within cities, discussed in “A city is not a tree” (Alexander 1965). There are echoes of the earlier preoccupation with the problem of morphogenesis, the synthesis of form, and in particular the mereological relation of parts and wholes, that has been Alexander's focus from the beginning of his career to this day.

The significance of A New Theory of Urban Design may be that it did, tentatively, propose a specific process by which a group of collaborators on an urban project might create such organic wholes more successfully, following a series of explicit rules.

In the experiment that forms the second section of the book, Alexander had 18 graduate students play various roles in a simulated process of urban growth. The roles included designers, developers, citizens, administrators, and others, making up a population not unlike that involved in the growth of a historic small city, or the community-led development of an actual neighborhood.

The role-players followed one overriding rule: every increment of construction must be made in such a way as to heal the city. Here he uses the word “heal” in the original sense of “to make whole.” That is, we must take a series of incremental steps in construction, and at each step we must make an assessment about whether the proposed construction adds to, or takes away from, the wholeness of the city.

He further defines this rule as follows: every act of construction has just one basic obligation: it must create a continuous structure of wholes around itself. That is, we must look at the way that the construction forms and changes patterns, and whether those patterns are patterns of coherent wholes, or of poorly related fragments and slivers. An extreme example of the latter is what planners derisively call “SLOAP” - “space left over after planning.” An example of the former is a nicely formed new spot that we would find appealing to be in, and that we would find well related to the spaces around it.

How are these wholes to be understood and manipulated? Alexander, ever the Cambridge mathematician, introduces a geometric entity he calls a “center” - a key element of the theory explained in more detail in his later and much larger work, The Nature of Order (2003-04). Philosophers of process will recognize strong similarities to Alfred North Whitehead's system of “actual entities” in that philosopher's magnum opus, Process and Reality (1928). But in essence, centers are simply localities, or “spots,” embedded within a field of other centers. A center is not a “point,” but rather, a field - as Alexander puts it, “a whole, made of subsidiary wholes.”

A field of centers, then, is a nested series of localities that frame one another and variously connect to one another in a pattern of relationships. Such a field comprises its own center at a larger scale. Conversely, every such center is embedded in a field of other centers that affect its structure and the structure of the wholes that result from their combinations. This field is contextual and infinite, but the contextual influence of more remote centers will generally recede with distance.

So too, every center incorporates other centers at smaller and smaller scales: houses, rooms, corners, floor tiles, patterns, etc. However, these smaller centers need not all fit neatly within the larger center: they can overlap with other larger centers as well.

The usefulness of this approach can be understood in a series of examples.

The cathedral design in Figure 3 can be seen as a relatively simple series of geometrically related centers. Every part of the structure feels coherent and well related to the other parts. Every region sits neatly nested within larger regions, with some occasional overlap, while every region is simultaneously composed of smaller regions. At the scale of a building, the plan feels whole and appealing. If it were a larger structure - if it were the plan for a city, for example - it might feel much too rigid and imposed.

Figure 4. Giambattista Nolli's engraved plan of Rome; a detail from his Nuova piata di Roma (1748). Note the much more complex system of messy, overlapping regions and centers.

Figure 4 shows a much more complex series of centers from Giambattista Nolli's engraved plan of Rome (1748). Highlighted are only a small number of them to illustrate the point. Note again that at the scale of buildings the centers are fairly neat and symmetrical. At larger scales, the centers tend to overlap and wend around one another, forming much more “organic” patterns. As Jacobs described, the “organic” character of these patterns is nothing other than the complex set of relationships by which they are “interrelated into an organic whole.”

Figure 5 shows an obvious counter-example. The large, undifferentiated geometry makes no attempt to adjust itself to its surrounding space. The shape of the paving, fencing, and other elements bears strikingly little relation to the space around them, and the overall form is unappealing in the extreme. This is a good example of the kind of outdoor space that planners derisively refer to as “SLOAP.”

The structuring process the students followed is as follows.

As one center 'X' is produced, so, simultaneously, other centers must also be produced, at three well-defined levels:

Larger than X. At least one other center must be produced at a scale larger than X, and in such a way that X is part of this larger center, and helps to support it.

The same size as X. Other centers must be produced at he same size as X, and adjacent to X, so that there is no “negative space” [i.e. SLOAP] left near X.

Smaller than X. Still other centers must be produced at a scale smaller than X, and in such a way that they help to support the existence of X.

Alexander and his students structure the centers of their projects with seven “detailed rules of growth.” Each of these is composed of still other sub-rules:

• Piecemeal growth governs the increment of growth, the distribution of elements, and the diversity of uses.
• The growth of larger wholes governs the way that emergent wholes are identified and reinforced, so that they become large-scale features of the urban plan. In this way, such large-scale features as main streets, public squares and the like emerge from the sequential process, not unlike the way they did in many historic cities.
• Visions ensure that the human participants will use their perceptual and imaginative ability at each step to assess the coherence of each increment. This is necessary to ensure that each increment is a well-adapted human space, with attributes of most value to human beings.
• Positive urban space ensures that there is no SLOAP around the structures, and that pedestrian space, gardens, streets, and other exterior spaces are well formed and coherent.
• The layout of large buildings provides detailed steps for the successful layout of area, size, entrance, main outdoor elements, major interior areas, and so on, for each large building. This section of the process comes very close to what Alexander later came to call a “generative code” (see below).
• Construction deals with the details of a building, and the iteration between how it is to be built and how it is to appear.
• Formation of centers provides detailed guidance of the geometric shaping of centers, as a kind of tension between its own internal tendency to create local symmetry, and the larger adaptation that tends to pull the symmetry apart. For Alexander the geometer, this is a characteristic feature of whole systems:

One of the reasons we can always recognize a real structure of centers as fast as we do is that we can always detect the truth in the balance of symmetry and asymmetry, even when we do not know what is going on functionally:

Thus, we may see the creation of the field of centers, as the creation of a loosely connected system of local symmetries, always relaxed, always allowing necessity to guide it, in such a way as to produce the deepest possible structure of centers, at every scale. (Alexander et al. 2003-04, p. 95)

Thus, Alexander's approach is fundamentally contextual, relying upon a continuous cyclical, stepwise response to existing conditions. In actual projects, Alexander often asks participants to walk the site, set stakes, meet residents, spend time, and let the site's attributes seep in (see Figures 6-9). He often continues this process through the design and the construction, even into the maintenance and repair of a completed project. In this exercise, Alexander and his students used a large-scale model as an approximation of the actual site. They did visit the actual site, and incorporated a number of its existing structures into the model at the outset.

Alexander's challenge

A New Theory of Urban Design amounted to a gauntlet thrown down to conventional urban design, not unlike that thrown down by Jacobs 26 years earlier.

Alexander himself was tentative about the particular methodology he proposed. Indeed, as discussed, he later offered his own critique of its shortcomings. But he was not then, nor has he been since, tentative about the key theoretical points on which this methodology differed from conventional practice:

• Urban design must not be an act of tabula rasa imposition of a form designed remotely, based upon an abstract program. It must understand, respect, and seek to improve the existing conditions.
• Urban design must incorporate the decisions and needs of the local stakeholders, as a matter not only of fairness, but also of the intrinsic quality of the result.
• Above all, urban design must be a generative process, from which a form will emerge - one that cannot be pre-planned or standardized, but will of necessity be, at least in some key respects, local and unique.

Implementation: the "many hands" of the New Urbanism

Perhaps the most notable example of an effort to implement Alexander's ideas - and Jacobs's in equal measure, it should be added - has been the New Urbanism movement. The Congress for the New Urbanism was formed by six architects, growing out of a 1991 workshop at the Ahwahnee Lodge in California's Yosemite Park. It was rather ironically modeled on the Congres International d'Architecture Moderne (CIAM), the highly effective organization that propagated the modernist movement in architecture, which accelerated the kind of segregation and top-down formalism in city-planning that both Jacobs and Alexander decried.

By contrast, the New Urbanism is explicitly about mixed use, and, its proponents would argue, about process.

By all accounts, Andres Duany has played a leading role in the creation of New Urbanism. His plan for Seaside, Florida, attracted enormous attention and prompted the coining of the term “new urbanism” by author Peter Katz (with Duany's encouragement). Duany credits Alexander as being a major influence on the New Urbanism, and has gone so far as to tell this author that Alexander's ideas are the basis of “everything that we're doing now” (Mehaffy 2004b). The New Urbanism includes a Charter with 27 principles for the structuring of urban form, including emphasis on mixed use, socio-economic diversity, historic preservation, walkability, and related objectives. It also includes a set of methodologies for the urban design process, at the heart of which is a workshop tool called a charrette.

Named after the cart that once gathered student drawings at the famous Ecole de Beaux-Arts in Paris, the charrette is an intensive design workshop that typically runs from four to nine days. It brings together urban designers, transportation planners, civil engineers, government officials, local residents, and other stakeholders and technical experts as the project may require. The Jacobsian and Alexandrian aim is to focus collaboration on design, and to produce an emergent result.

It is important to understand that a charrette is not merely a “user consultation,” of the sort that was commonplace before The New Urbanism, but a real-time design process in which users provide collaborative input, along with an interdisciplinary mix of professionals. It is, according to Duany, an exercise to provide exactly the kind of collaborative synthesis that Alexander described in A New Theory of Urban Design.

Alexander's critique

For Alexander, however, the charrette is a laudable effort at reform that is still woefully inadequate for the challenge. His criticism rests on three principal objections:

• The charrette process is still a relatively brief and isolated act of master planning done in a remote room, away from the site, and away from the opportunities and constraints that might turn up in a longer and more direct process of contextual engagement.
• The participants, especially the local residents, are forced to play a highly circumscribed role, in which the “outside experts” disproportionately influence the process.
• The master plan is then usually turned over to a developer, who can then interpret the master plan in a variety of usually disastrous ways. Most importantly for Alexander, the developer typically builds structures that are not at all generative, but based upon standardized templates, with the result that they feel lifeless and unsuccessful. They may have the outward appearance of a more organic neighborhood, but they are, in the end, standardized reproductions.

For Alexander, the most serious problem is the fact that the output of the charrette - a “master plan” - is usually turned over to developers:

I think that many of the people who are involved in the CNU actually have not understood the problems that the developer represents, and what has got to be done in order to change that situation. It's very very serious. (Mehaffy 2004a)

Alexander clearly feels great sympathy with the New Urbanists, but equally clearly feels unease at the result of New Urbanist work:

I'm proud of them, because they've really done something to help change things. But when you say, well, what are the rules that they actually live by? I'm talking about “live by” when they're shaping something, modeling it, drawing it, planning it, things like that - building it, and so forth - the concepts that they are living by there are not those which I've just been speaking about, having to do with whether you're making part of nature. They're actually something highly artificial …. (Mehaffy 2004a)

The New Urbanists' rebuttal - and a counter-critique

For New Urbanists such as Duany, Alexander's critique misses a key point. Yes, there are standardized templates within The New Urbanism - as, for example, the so-called “parti,” a basic plan drawing of the scheme. But that structure can then be adapted and allowed to serve as a skeletal form for more organic growth. In effect, it can serve as a kind of well-designed “trellis” on which organic growth can self-organize. Duany notes that such combinations of the standard and the contextual are common in nature.

Duany and others point to Alexander's own patterns as typological structures that are, in part, standardized elements within his own design system (though a networked one, and not a strict hierarchy). They are then adapted to the specific context, and used in a kind of flexible grammar. Duany believes he is doing something very similar (and indeed, often using Alexander's own patterns). “I am the best Alexandrian,” he recently told the author.

Moreover, Duany believes Alexander is failing to come to terms with a core reality of modern technological society:6 that large numbers demand top-down management methods. In a mass society, the norm quickly reverts to chaos and kitsch. In order to implement Alexander's methods, this demands expert, top-down leaders for the design and construction process - a role, Duany points out, that Alexander himself often plays in his own projects. But the scale of reform does not permit the kind of painstaking one-off approach for which Alexander is known.

These points were evident in comments by Duany in an interview with the author (Mehaffy 2004b). He was asked to expand on his comment that “getting things implemented on a large scale” was one of the aspects of modernity that interested him:

Isn't there a danger, as Chris Alexander has warned, that that kind of large scale doesn't allow the grain of adaptation that's going to be required for good urbanism and good architecture? Isn't scale one of the key things that the modernists screwed up on so badly?

Duany replied:

Yes scale is a prime problem, but it is also THE reality of modernity. It cannot be avoided. I believe it was Giedion, the theoretician of modernist architecture who said: “Ours is the problem of large numbers”. These early modernists were very smart. There is immense population growth, even in the U.S.; there is the blinding speed of communication, of automatic decision protocols and of mass production. If we return only to the crafting of cities, what we do may be of high quality but it will not be important. (Mehaffy 2004b)

For Duany and other critics, Alexander's proposal is to return to a painstaking one-off process of organic design, which is simply not up to the scale of the present challenge. Rather, we must create more automatic processes that generate the same result, not unlike seeds that generate vast numbers of living structures:

… What we must craft now are not the communities, but the programs that create them quickly automatically, replicating the organic process of sequential decisions. This must achieve the authentic variety and resilience of traditional communities. These protocols must be propelled by the power grids, and not be tied to the limits of our personal efforts. Every one of our “adversaries” operates from automated protocols. Protocols that flow 200 mortgages for 200 parametric strip shopping centers, all bundled for purchase in a single transaction by an insurance company. What would they do with our one-off creations? Inspect them individually? They can't. (Mehaffy 2004b)

The reference to “power grids” echoes a discussion that Duany and Alexander had in 1988. According to Duany, Alexander said to him, “we both know what the appliance is. What we need to do now is to design the plugs to connect to the current power grid” (Mehaffy 2004b). For Duany, Alexander has neglected this task, whereas the New Urbanists have pursued it with full force - accounting for the latter's much more prodigious output of projects.

So Duany and other New Urbanists have turned to a new project: the development of codes that replace the old, destructive protocols with new ones that allow good urbanism to flourish, as if on well-constructed trellises. The “SmartCode” is a form-based code that replaces the segregated “Euclidean” zoning of an earlier era with a series of parametric specifications designed to ensure coherent streetscapes and public realms. The code uses a “transect” system to organize contextual responses to the urban condition, from the most intense urban setting to the most pristine natural environment.

Alexander's "generative" code

But for Alexander, again, this kind of code does not address the core prerequisite of generativity, and without such guidance for growth the result is still likely to be well-aligned, lifeless junk. It prescribes a series of static parameters within which generative events may occur, but it does not in any way facilitate or guide their generation. Moreover, even to specify such parameters is to constrain the emergence of organic wholes, which require an environment in which adaptation of form can occur as needed.

Almost in response to the New Urbanists, it would seem, Alexander has proposed an alternate kind of code, based explicitly upon rule-based, generative processes of the kind outlined in A New Theory of Urban Design. Alexander's “generative code” addresses not physical parameters of the built environment, but steps that the participants should take together in laying out and detailing a given structure. Alexander likens it to a recipe, or a medical procedure, in which the steps always follow a logically similar pattern, but the actual actions continuously adapt to the context - the taste and texture of the food in the case of a recipe, or the condition of the patient's tissues in a medical procedure. But in this case, the “recipe” or the “procedure” guides the unfolding of environmental form.

In its fullest form, this kind of generative code can be thought of as a design-build system, addressing all of the conditions of building - financing, ownership, management, sourcing, and, crucially, changes to the design along the way. For Alexander, the issue of cost control is a manageable process, and indeed, is done regularly within existing design-build approaches. He points out that much of the direction of technology is today aimed favorably for such an approach - one-off manufacturing, customization, niche marketing, and so on. He is convinced of the possibility and even the inevitability of this transformation of technology, in a more adaptive, ultimately organic direction.

Nonetheless, Alexander recognizes that there are enormous challenges ahead to making a practical version of such a system. He continues to work with a growing group of collaborators (including the author) on such a project, and he has repeatedly stated that he welcomes the opportunity to develop collaborations with New Urbanists like Duany, as well as others.

Alternative approaches to generativity

Duany's discussion of the “problem of large numbers” would find a sympathetic audience with the architect and theorist Rem Koolhaas, his forensic opponent in a rather lackluster debate in 1999 at Harvard's Graduate School of Design. For Koolhaas - perhaps representing many other contemporary “neo-modernist” architects - the modern city is simply too complex to yield to a reform agenda like that of the New Urbanists. In the face of sheer quantity, architecture is powerless to change the direction of the urban wave, and therefore is wiser to seek merely to surf that wave with skill:

This century has been a losing battle with the issue of quantity. … In spite of its early promise, its frequent bravery, urbanism has been unable to invent and implement at the scale demanded by its apocalyptic demographics. (Koolhaas 1995)

Koolhaas challenges Duany's faith in planning, and suggests that urbanism is now the art of accommodating generativity, rather than the futile attempt to “design” it:

If there is to be a “new urbanism” it will not be based on the twin fantasies of order and omnipotence; it will be the staging of uncertainty; it will no longer be concerned with the arrangement of more or less permanent objects but with the irrigation of territories with potential; it will no longer aim for stable configurations but for the creation of enabling fields that accommodate processes that refuse to be crystallized into definitive form …. (Koolhaas 1995)

Another approach to generativity is typified by Peter Eisenman, Alexander's partner in a famous and telling debate in 1982, since billed as “Contrasting Concepts of Harmony in Architecture” (Eisenman and Alexander 1982). In his book Code X: The City of Culture of Galicia (Eisenman Architects 2005), Eisenman discusses his theory of coding as a generative method of producing form. The theory was put to the test in the City of Culture of Galicia project in Santiago de Compostela, Spain, the most urban in scale of Eisenman's projects (see Figures 10-13).

The project narrative describes a transition from an architecture of semiotics (symbolic expressions) to an architecture of generated geometries or “traces:”

This post-semiotic sensibility is not dominated by easily consumed imagery of signs and signifieds, but rather is understood as a series of traces, marks that produce an alternative condition of figure and ground. The City of Culture evolves from the superposition of three sets of traces. First, the plan of the medieval center of Santiago is placed on the hillside site, which overlooks the city. Second, a Cartesian grid is laid over these medieval routes. Third, the topography of the hillside is allowed to distort the two flat geometries, thus generating a topological surface that superposes old and new in a simultaneous matrix.

The original center of Santiago conforms to a figure/ground urbanism. The buildings are figural and the streets, residual. Through this transformative mapping operation, our project emerges as a warped surface that is neither figure nor ground but both a figured ground and a figured figure that supercedes the figure-ground urbanism of the old city. Santiago's medieval past appears not as a form of representational nostalgia but as an active present found in a tactile, pulsating new form. (Eisenman Associates 2007)

The generativity in Eisenman's approach can be contrasted with the generativity exhibited in the nearby historic town it echoes (Figure 10). The latter has emerged from the collaborative and rule-based actions of many actors over time, and from the environmental conditions to which they have adapted. Artistic abstractions and planning schemata do occur, but are expressed as local elements within the more global adaptations to environment, culture, human activity and need.

By contrast, Eisenman's generativity is used solely as a resource for one artist's expressive master plan, imposed on the site at a very large scale. In that sense its semiotics is in fact alive and well, but disguised within a subtler artistic reference to incidentally generated traces of its natural subject. It regenerates only the most skeletally abstract aspects of the historic evolutionary pattern, so as to avoid “representational nostalgia.” It is otherwise a static and non-adaptive work of art.

Koolhaas's and Eisenman's positions here can be contrasted with Jacobs's.7 For Jacobs, urban practice was a proper intervention in the interest of the health of an urban system, accomplished by patient inductive study and by manipulation of subtle catalytic factors. Art was a dimension of this work, but far from its only dimension. She would arguably regard Koolhaas's nihilism as little more than the predictably frustrated reaction to a continued failure to adopt the most recent and most accurate model of “the kind of problem a city is.” She would arguably regard Eisenman's position as an altogether different model - a hijacking of the city by fine artists, who would see it transformed into an enormous abstract sculpture gallery. This, she frequently warned, was a dangerous attitude: “The city cannot be a work of art” (Jacobs 1961, p. 372).

Interestingly, both Koolhaas and Eisenman have recently expressed remarkably pointed misgivings about the efficacy of their approach, and the larger artistic culture of which it is part. Speaking in Montreal in June 2007, Koolhaas lamented the effect of laissez-faire market forces on the profession:

If you look back over the past 2,000 years, architecture dignified civic and public life. Then the market economy happened, replacing all former values and erasing almost all ideology over the entire world. What it represents for all of us today is an invitation to simply be extravagant and spectacular. … The work we do is no longer mutually reinforcing, but I would say that any accumulation is counterproductive, to the point that each new addition reduces the sum's value. … So there are many problems, first of all our work, which is not able to find its way out of this recurring dilemma, then there are the many reasons to question our sincerity and motives. (quoted in La Giorgia 2007)

Eisenman, speaking at the same event, argued that we are in the late period of modernism - its “death rattle” - but we are struggling to find a new paradigm to replace it:

We are in the rococo phase of modern architecture. … The problem we need to solve is the urgency of media to have something new to look at and talk about all the time. Our need to be in the news all the time. … The slowness required to find and understand meaning in architecture no longer has any attraction. [We need an architecture] that asks how, at this moment in time, without a new paradigm, can we understand our discipline and our culture in a different way. (quoted in La Giorgia 2007)

Or perhaps we need to look more deeply for a new paradigm within the insights of modern science and philosophy. This is precisely what Alexander has said he is seeking.8

In the planning disciplines, generativity has continued to develop in the work of other investigators. In particular, the trend toward engagement of residents evident in the “third generation” of the design methods movement has continued and accelerated. A notable example is so-called Communicative Planning, which seeks to build inclusiveness, incorporate difference, reach out to marginalized groups, and sensitize planners to a wide variety of viewpoints and alternative ways of knowing (Qadeer 1997, Sandercock 2000, Harwood 2005). Sandercock (1998) describes the evolution of a “utopia with a difference” (pp. 5, 119) a similar concept to Friedmann's (2002) “open city” of diverse peoples, united by applied principles of ecology, citizenship, and regional governance. The planning profession has been plagued by the lack of a workable knowledge base about how to communicate with diverse population groups (Wallace and Milroy 1999), but we now see how “a thousand tiny empowerments” can help to constitute a more socially transformative planning process (Sandercock 2000).

Communicative planning seeks to achieve collaborative consensus-building by, in effect, developing “'conversations' between stakeholders from different social worlds” (Healey 1997, p. 219). Innovation, “drama,” and “a sense of play” are ways to “move the players and embed their learning deeply” (Innes and Booher 1999, p. 19).

Conclusion

For all their disagreements, the cross-fertilizations between Alexander's process advocates and the New Urbanists continue, with constructive results. The topic of generativity continues to loom large.9 Duany's SmartCode - now adopted by dozens of municipalities and under consideration as the national planning code of Scotland, among others - has begun to take on some stepwise layout guides very similar to Alexander's. (Some Alexander allies, including this author, continue to urge the expansion of this offering.) Duany argues that his code also incorporates many other aspects of generativity. For his part, Alexander has continued to develop his proposal for a “generative code,” and to address the “massive process difficulties” that are posed by conventional building protocols, using many of the New Urbanists' insights.10

To be sure, Alexander faces daunting challenges that the more pragmatic New Urbanists seem uniquely positioned to help meet. In fact, each seems to have a complementary grasp on aspects of the problem that the other, through area of focus or through sheer personality, seems much less able to address. This emerging model of collaboration may hold more promise than either may realize.

A growing group of collaborators has assembled around this sharable, complementary agenda, and begun to pursue lines that Alexander (for one) does not seem to find as interesting - most significantly, “open-source” collaborations with biologists, ecologists, sociologists, computer scientists and others. Such “open source” methods have yielded remarkable results for the computer software developers who exported Alexander's ideas into that realm with remarkable effect.11

In an age of critical ecological and economic challenges, in which human technology seems at nearly irreconcilable odds with ecological sustainability, Alexander argues that we must have a much more serious look at the way that natural systems use generative processes to achieve sustainable morphologies, and work to integrate those lessons into our own human systems.12 Though progress has been slow - and yet, as has been argued herein, substantial - Jacobs and Alexander demonstrate that this is a comprehensible problem, and not one that is (to quote from Jacobs's caricature, paraphrasing Warren Weaver) “in some dark and foreboding way, irrational.” The opportunity remains to develop further generative processes as a means to deliver more robust and more efficacious results - that is, more sustainable results - within the field of urban design. But that task will surely demand the combined and synergetic efforts of Alexander, Duany, and many others.

Notes

1. A “design parti” was a schematic diagram used early in the Beaux-Arts design process. “Parti” means to divide, hence to organize basic regions of the plan schematically in a diagrammatic scheme.

2. Note, however, that Alexander does use master plan drawings as a form of design communication, or entitlement documentation. He is, however, careful to emphasize that they are snapshots in a longer process, and not any sort of “final” result. He does so by including explicit generative processes as part of the planning documents. See, for example, The Master Plan and Process for Harbor Peak (Alexander 2006: http://www.livingneighborhoods.org/library/brook-1.pdf).

3. Several correspondents have related discussions with Jacobs along these lines. The author is particularly indebted to Arun Jain, Chief Urban Designer for the City of Portland, who discussed these matters with Jacobs on a number of occasions.

4. Reported in a seminar discussion at University College London, where the author was present.

5. Most recently, the author spotted a well-thumbed copy sitting conspicuously on the shelf of the Executive Director of a major New Orleans preservation charity. As it was pointed out to her, she remarked, “Oh, yes - I love that book!”

6. The author is indebted to Andres Duany for a number of conversations on this topic. Any errors in representing his views are entirely the author's own, for which apologies are given in advance. See in particular his interview in Mehaffy (2004b).

7. At any rate, Jacobs did seem to regard the urban interventions of Koolhaas's contemporaries, including Eisenman, with dismay. In a letter to the author in 2001, she related that she was “appalled” at the proposals for Ground Zero in New York. She only refrained from getting involved, she said, because she was no longer a New Yorker. But she referred the author to other colleagues in New York who were said to be preparing to oppose the plans.

8. There is an extensive discussion of this topic in Grabow (1983). Alexander also discussed this topic in the present author's interview with him (Mehaffy 2004a, 2007).

9. Indeed, even at the time of writing, it is the subject of a very animated exchange on a New Urbanist listserv, including Duany, his colleague Sandy Sorlien, the present author, and others.

10. At the Congress for the New Urbanism in 2006, Alexander held a meeting at which some 30 people, including a number of prominent developers, pledged to collaborate with him. A listserv was formed, and plans were made for a symposium - which was put on hold when Alexander was unable to finalize an agreement with the Prince's Foundation for the Built Environment to host it.

11. See, for example, the explosive growth of the “design pattern” movement in software, based on A Pattern Language, and begun by former Tektronix engineers Ward Cunningham and Kent Beck. Alexander's ideas have thus directly spawned the development of such familiar titles as Wikipedia and The Sims. Cunningham has been involved in the more recent collaborations to develop Alexander's ideas on generativity further.

12. For example, very promising and hopeful work is being done within game theory and economics, notably in the realm that seeks to integrate so-called “externalities” within more sustainable economic processes. This echoes Alexander's efforts to “change the rules of the game” of real estate development.

References

• 1. Alexander, C. (1964) Notes on the synthesis of form Harvard University Press , Cambridge, MA
• 2. Alexander, C. (1965) A city is not a tree [2 parts].. Architectural Forum 122:(1) — 58-62 (Part I); 122 (2), 58-62 (Part II). Available from: http://www.arq.ufmg.br/rcesar/alex/alexander/alexander1.html/
• 3. Alexander, C. (1979) The timeless way of building Oxford University Press , New York, NY
• 4. Alexander, C. (2003-04) The nature of order, Books 1-4 Center for Environmental Structure , Berkeley, CA
• 5. Alexander, C. (2006) The master plan and process for Harbor Peak: a model creation process for 21st century cities Center for Environmental Structure , Berkeley, CA
• 6. Alexander, C.,  Ishikawa, S.,  Silverstein, M.,  Jacobson, M.,  Fiksdahl-King, I. and Angel, S. (1977) A pattern language Oxford University Press , New York, NY
• 7. Alexander, C.,  Neis, H.,  Anninou, A. and King, I. (1987) A new theory of urban design Oxford University Press , New York, NY
• 8. Batty, M. (1991) Cities and fractals: simulating growth and form Springer , New York, NY
• 9. Batty, M.,  Couclelis, M. and Eichen, M. (1997) Urban systems as cellular automata Pion , London (UK)
• 10. Broadbent, G. (1979) The development of design methods a review.. Design Methods and Theories 13:(1) , pp. 41-45.
• 11. Dulgeroglu-Vuksel, Y. (1999) Design methods theory and its implementation for architectural studies.. Design Method: Theories, Research, Education and Practice 33:(3) , pp. 2870-2877.
• 12. Eisenman Architects with Davidson, C. (ed) (2005) Code X: the city of culture of Galicia Monacelli , New York, NY
• 13. Eisenman Associates, 2007. Narrative for design of City of Culture, Santiago de Compostela, Spain [online]. Columbia University Graduate School of Architecture, Planning and Preservation. Available from: http://www.arch.columbia.edu/index.php?pageData = 32552/
• 14. Eisenman P, Alexander C., 1982. Contrasting views of harmony in architecture. Transcript of a debate at Harvard Graduate School of Design [online]. Katarxis 3. Available from: http://www.katarxis3.com/
• 15. Fowles, B. (1977) What happened to design methods in architectural education?. Design Methods and Theories 11:(1) , pp. 17-31.
• 16. Friedmann, J. (2002) City of fear or open city?. Journal of the American Planning Association 68:(3) , pp. 237-243.
• 17. Grabow, S. (1983) Christopher Alexander: the search for a new paradigm in architecture Oriel , Boston, MA
• 18. Harwood, S. A. (2005) Struggling to embrace difference in land-use decision making in multicultural communities.. Planning, Practice and Research 20:(4) , pp. 355-371.
• 19. Healey, P. (1997) Collaborative planning - shaping places in fragmented societies Macmillan , London (UK)
• 20. Hillier, B. (1989) The social logic of space Cambridge University Press , London (UK)
• 21. Hillier, B. (1999) Space is the machine: a configurational theory of architecture Cambridge University Press , London (UK)
• 22. Innes, J. E. and Booher, D. E. (1999) Consensus building as role playing and bricolage: toward a theory of collaborative planning.. Journal of the American Planning Association 65:(1) , pp. 9-26.
• 23. Jacobs, J. (1961) The death and life of great American cities Vintage , New York, NY
• 24. Koolhaas, R. Koolhaas, R. (ed) (1995) Whatever happened to urbanism.. S, M, L, XL pp. 958-971. Monacelli , New York, NY
• 25. La Giorgia G., 2007. Market v. meaning. Architecture Week. 5 September [online]. Available from: http://www.architectureweek.com/2007/0905/design_3-1.html/
• 26. Mehaffy M., 2004a. A conversation with Christopher Alexander. Katarxis 3. Available from: http://www.katarxis3.com/
• 27. Mehaffy M., 2004b. An interview with Andres Duany. Katarxis 3. Available from: http://www.katarxis3.com/
• 28. Mehaffy, M. (2007) On the nature of order: an interview with Christopher Alexander.. Urban Design International 12 , pp. 51-57.
• 29. Qadeer, M. A. (1997) Pluralistic planning for multicultural cities: the Canadian practice.. Journal of the American Planning Association 63:(4) , pp. 481-494.
• 30. Sandercock, L. (1998) Towards cosmopolis Wiley , Chichester (UK)
• 31. Sandercock, L. (2000) Cities of (in)difference and the challenge for planning.. Dokumente und Informationen zur Schweizerischen Orts-Regional und Landesplanung (DISP) 140 , pp. 7-15.
• 32. Saunders W., 2002. Book Review: A Pattern Language. Architectural Record. May, 93. Available from: http://www.gsd.harvard.edu/research/publications/hdm/back/16books_saunders.html/
• 33. Simon, H. (1962) The architecture of complexity.. Proceedings of the American Philosophical Association 106 , pp. 467-482.
• 34. Wallace, M. and Milroy, B. Fenster, T. (ed) (1999) Intersecting claims: planning in Canada's cities.. Gender, planning and human rights pp. 55-73. Routledge , London (UK) — In
• 35. Whitehead, A. N. (1928) Process and reality Free Press , New York, NY

List of Figures

Figure 4. Giambattista Nolli's engraved plan of Rome; a detail from his Nuova piata di Roma (1748). Note the much more complex system of messy, overlapping regions and centers.