CECIL BALMOND has made a career out of doing what he’s not supposed to. Trained as an engineer, Balmond has radically expanded the traditional role of that profession, building a reputation as one of the world’s leading structural designers. Over the past four decades, he has had a hand in shaping many of the world’s most significant buildings—working with renowned architects from James Stirling and Philip Johnson to Rem Koolhaas and Toyo Ito—and has collaborated on major public commissions with artists such as Anish Kapoor. The two constants underlying this extraordinary diversity of projects are Balmond’s unique spatial sensibility and his unparalleled mastery of new digital technologies, which are now the driving forces behind Balmond Studio. Founded in 2010, the firm pursues cutting-edge research in design and computation while producing commissions in both art and architecture. Artforum invited Balmond to speak with senior editor JULIAN ROSE about the genesis and trajectory of this sweeping—and continually surprising—body of work.
JULIAN ROSE: It’s an unfortunate paradox in the history of modernism: Even as new technologies have become more important to the practice of both art and architecture, technical concerns tend to remain isolated—a set of concrete, real-world problems may need to be solved for a work to be realized, but they remain separate from that work’s more ineffable aesthetic or conceptual significance. You have been able to obviate such distinctions. Indeed, early in your career, you gained a reputation as an exceptionally innovative structural engineer, but in many ways you have always been an artist. And so your work with architects has been characterized by intense partnerships that are closer to coauthorship than the typical structure of an engineer working for, rather than with, an architect; you have also designed architectural projects independently and, more recently, extended your activities into the realm of art, founding a studio that produces everything from installations and exhibitions to public sculptures and computation research. How did this sprawling practice come about?
CECIL BALMOND: Well, from the beginning, I didn’t believe the world was as simple as it was typically made out to be. In the 1970s, when I was a young structural designer at the London-based engineering firm Arup, the buildings we worked on always seemed to be subject to a reductive categorization, broken down into a discrete set of parts—columns, slabs, and beams. I was uneasy with this thinking, which seemed to be a consequence of modernism and its stripped-down measures. Design discussions would focus only on the size of the members and the spacing of the column grid. An architect who wanted to minimize the visual impact of the column in a space would simply ask to make it smaller and smaller, as if it would eventually vanish. And the engineer would just focus on doing that—making the column as efficient as possible, using the minimum amount of material required. But no one would ever think of actually moving it out of sequence, or perhaps even thickening it, to provoke a spatial question. No one was interrogating the underlying order.
I felt that this standard response was extremely limiting in terms of the architecture it could produce. I wasn’t interested in the column just as a support; for me, it was a punctuation point in space. I didn’t see structure as a skeleton. I saw it as a periodic system that created architectural episodes, events—a temporal experience. So I became restless. I wanted to move structure around, to question assumptions, and to see what spatial possibilities could come from that—it was the articulation of space that interested me.
JR: The Maison à Bordeaux [Bordeaux Villa, 1998], which you worked on with Rem Koolhaas, is an impressive early example of how you were able to manipulate structure and space—it changed our way of thinking about support.
CB: Generally, buildings are made from a post-and-beam configuration, which is thousands of years old. When they were using stone to build like this in ancient Greece, the structure had to be precisely symmetrical and balanced, with the columns directly below the load, because stone is relatively weak. Today, we’ve got reinforced concrete that can transfer loads in all kinds of ways—you can do almost anything with this stuff—but nobody had ever really tried.
The design for the Bordeaux Villa changed that. It questioned the assumption that a column has to be right underneath the weight it supports. I took the heavy mass of the main building and broke with the symmetry you would expect from the structure, pushing the columns around, out to the side. I replaced a classical order with an eccentric geometry. So when you look at the villa, you feel like it’s floating. Rem and other architects started to work with me because they saw I was attempting to shift the boundaries of architecture—literally. Structure was no longer a constraint, and architects found that very liberating. Structure didn’t even have to enter the discussion. We could just talk about form and space, rhythm and texture.
So after many collaborations, I went into architecture myself. Fourteen years ago, I founded the Advanced Geometry Unit [AGU] within Arup. This was a multidisciplinary design-research group—with architects, specialist engineers, computer scientists, a game theorist, even a quantum scientist. I had begun to think of space as unknowable—the opposite of the rational space of the modernist grid. In other words, I didn’t want to nail down space and structure or provide a singular, rigid solution, which had been the traditional engineering approach. Why couldn’t architecture be more fluid, improvisational, relative? How does space manifest itself through the intervention of mass? What are the material or structural moments that catalyze new readings of space? To explore these possibilities, the AGU abandoned conventional rule-of-thumb calculations and design methods based on proportion or repetition and began to seek out new, more abstract, organizational systems. We deployed these approaches in architectural projects; one was a footbridge in Coimbra, Portugal, which was completed in 2006. But we were getting into new ways of thinking about form, space, composition, and many other things—with implications far beyond architecture.
JR: And this is when you moved toward art?
CB: Yes, with the creation of Balmond Studio four years ago. The studio is made up of a group of designers with diverse backgrounds, some having worked in architecture and some in art. The emphasis is on an ideas laboratory, with research being a key part of the studio.
JR: Although it’s not uncommon these days for artists to have large studios, the model of production is often relatively traditional, still based on the role of the studio assistant, really, where certain technical problems—fabrication, say—are outsourced to various experts or specialists. It sounds as though you’re trying to create a fundamentally different mode of working.
CB: Yes—I think there’s often a dumbing down of fabrication today, where the norm is simply to bring industrial processes directly into art production, and artists miss out on the potential that comes from genuine collaboration on technical issues. I use a fabrication outlet in Nottingham, UK, where I prototype the work, and I challenge my fabricators to leverage their expertise in materials and construction to bring out aesthetic solutions to the problems posed by a given work. We’re always building a range of options, testing material solutions against the conceptual content of a piece. For a recent public sculpture [Snow Words, 2012] in Anchorage, Alaska, for example, we played with patterns of interference to produce extremely rich and subtle lighting effects from LEDs, and I’m sure if we had simply hired a lighting expert to do the installation, this would never have occurred to us.
But the studio’s structure isn’t so much about working in one field or another as it is about moving beyond reductive thinking and breaking away from singular categories and assumed definitions, whether we’re talking about a building or a sculpture. I’m interested in a more interactive, integrated approach, based on my belief that the world is irreducibly complex.
JR: In fact, your work seems to take complexity—formal, spatial, and visual—as its very raison d’être. Thinking in terms of such intricate and interconnected systems—where a column, say, is not a fixed point of support but one variable in a much broader network—requires introducing an extraordinary quantity of information into your design process. Handling enormous amounts of information is something computers are very good at, and, not surprisingly, you were also a pioneer of computational design. How have digital tools influenced the development of your work, and even your negotiation between fields?
CB: There was an important turning point in my use of the computer in the early 2000s. When I collaborated with Anish Kapoor on Marsyas, a sculpture for the Tate Modern’s Turbine Hall in 2002, the computer was seen as a tool—very powerful, but just a tool. The form of the piece, an enormous tubular skin with three openings, could have been produced without the computer. The problem was getting that shape to be completely smooth, so that the form would take on the feel of solid material—we wanted it to look and feel almost like steel. The fabric was stretched between three rings. As you pull the membrane tight, the forces of tension constantly change. So you need the computer to run a series of iterative simulations until you arrive at a pattern that distributes the forces evenly and allows the surface to be taut, wrinkle-free. The computer’s power is invaluable for this kind of trial-and-error process. There are so many variables changing that you could never really figure it out without the computer. If we had tried to just hang the piece without running these sophisticated programs, it would have been wrinkled like crazy. And normally a fabric structure would have cables and masts attached to it, to hold its shape. In Marsyas, the orientation of the three rings forced the fabric to shape itself, flying through Tate Modern, defying gravity. So this was a major achievement, formally and structurally.
But around that same time, my team at the AGU began coding more abstract approaches. We would write software based on an intuition I might have had about a specific rule set leading to the development of a form, much as a tree has the command “branch” throughout, from its roots to its trunk to its leaf pattern. Form was no longer static—we were producing it by repeating a base idea, layering, augmenting. This was a whole new ball game. Computers became intelligent; writing and running the software became an act of discovery. By the time I worked with Toyo Ito on his design for the 2002 Serpentine Pavilion, I was using the computer in a totally different way. Ito and I had talked about the building having a random roof, with beams crisscrossing all over the place. But it would have been held up by a relatively typical array of columns, which just didn’t look right to me. Besides, total randomness is usually quite boring. It’s more interesting to find some kind of hidden order, present but not explicit to the viewer.
So the problem became one of producing a system that had an element of disorder or eccentricity but that could also become the building itself, that wouldn’t need a secondary system to hold it up. To solve this, a simple rule was used—an algorithm. I began by drawing a line from the halfway point of one side of a square to a point one-third of the way along the adjacent side. The rule was repeated many times, creating a network. Eventually I had a kind of mesh that I could cut and fold and turn into the building, materialized by translating lines on paper into trajectories of steel. And we didn’t have to worry about efficiency in the traditional sense of establishing a structural hierarchy. We had a multitude of intersecting lines, and we could use them all—the entire network became a structure of mutual support. It was a surprisingly economic solution, too— actually more efficient than using a traditional framing system for the roof.
If Marsyas was quite willful—we had a form in mind, and the computer was used as a high-powered shaper to make it smooth—with the Serpentine Pavilion there was no preconceived shape for the building, only a desire for crossed lines. Then the computer’s intelligence helped me produce the network that became the building. This totally upended the normal design process. It’s the first example, I think, of an algorithm genuinely being used to define an architecture in its entirety.
JR: But is there a danger in assigning the computer too much of a generative role? I’m curious to hear more about how you apply these algorithmic tools in your own artistic practice. Because there’s a long history in art, and architecture as well, of using technology to suppress or subvert authorship—think of the prevalence of industrial fabrication techniques in Minimalism and post-Minimalism. And today, two of the biggest myths surrounding computational design are so-called optimization and emergence. The first is the idea that there is a single, ideal solution that the computer can produce, and the second is the belief that the designer fades into the background, letting the “best” design emerge automatically from the code. This is good salesmanship, as it can make even the most eccentric designs seem objective and rigorous. But it would also seem to grossly oversimplify the implementation of these digital tools. How do you understand your role as an artist and author when using the computer?
CB: It’s true that people today put too much faith in the idea that the computer will offer a singular solution. I see a lot of young students, in particular, who think that if they write an algorithm it’s the answer to everything.Well, it’s impossible to “solve” a whole project with one algorithm. Even if the computer is an intelligent tool, you still have to use your intuition, and you also have to be aware of the limitations of the technological system you’re using. Actually, if you look around, computational processes are involved in almost all aspects of our daily life. They may be intelligent, but they’re not human, and they tend toward homogeneity, which is dangerous. You have to know when to disrupt the process. There are certain intangibles that are totally human—breaks, interruptions, failures—that are extremely important to a designer; they are part of your role as an author. The system can’t be frictionless.
JR: Is this the model behind your most recent projects, where you seek to complicate or destabilize the computational system in some way?
CB: Yes. To give you an example, I’m currently working on a project called Sigma, which is a series of algorithms that interact to produce surprisingly beautiful forms. My starting point was the realization that when most people talk about algorithmic design, they are really only using a single algorithm. They take a data set, they run it through a single process, and something falls out of it. That’s what I did with the Serpentine Pavilion, for example. But I wondered what would happen if there were multiple algorithms that nested within one another. I wanted to allow breakage or failure. I wanted to ask, How complex is complex?
Thinking of the four components that make up the DNA molecule, I designed four algorithms that interrelate and interact in abstract space, not in a helical spiral but in two opposing vortices.At first, I was exploring it very primitively, by hand, just looking at the way a series of points were being connected in space. A lot of algorithms used in statistics or encrypted code are mathematical: The ones used in Sigma are essentially spatial, multiplying growth sequences. They are qualitative, although still attached to quantitative inputs. And people often treat an algorithm as if it’s some kind of magical abstraction, a black box, but really at first it can be incredibly basic, essentially just setting up sequences of movement between points in space. You don’t even need a computer; that comes later. Getting to know an algorithm, you’re just feeling your way; it’s more shamanism than science.
Once you have an intuition about how the systems interact and what kind of input you want to give them, that’s where the computer comes in. I’ve had an idea about trying something like this for maybe fifteen or twenty years, but it’s only within the past few years that computing has gotten powerful enough for me to try it. While the algorithms are not so complicated, it’s the computer’s ability to run these relatively simple operations iteratively, in a nested manner, and in three dimensions (or more) that is important. As with DNA, even if the rules by which I’m connecting points are easy to grasp, by running many successive operations and recombinations, I can very quickly get to something that is far too intricate to be done by hand.
The visualizations provided by the computer can also initiate a feedback loop. You can pull some images out of the computer and then go back and tweak the data that you’re entering, and you start to get an intuitive sense of what the system can do, the patterns it produces and the forms it assembles. The most interesting thing to me about Sigma is that it gives back both organic-looking objects and highly crystalline networks. The project was never intended to be about biomimicry, but there seems to be a natural-selection process at work.
JR: And what about time? An algorithm, after all, is a process—what is the time frame in which this program unfolds and in which you are interacting with it?
CB: Yes, most people forget that an algorithm is an ongoing operation, so there is no one answer. That’s very exciting, but it also presents a problem: It can always keep going, so where do you stop? With Sigma, I find that after about eighteen iterations, a certain formal complexity arises. As I’m watching the data emerge and converge, that’s the point where it seems like the systems are intelligently talking to one another. On the other hand, when the algorithm is poor or the data set is loose, with no inner relationship, you might run up to one or two hundred iterations, and it will remain chaotic and meaningless—you’re just picking up noise, not getting anywhere.
So an algorithm is a process that you set in motion, but there are still many judgment calls that you have to make as an author. Not only is there an intuition about how to start it, but then there is a series of decisions about when to stop, when to rerun it, what to plug into it. It would be a mistake to think that there is some Frankenstein machine that you can go to, switch it on, and your next design will pop out. What is new in this work is that the real design is the design of abstraction, the design of thinking, which is followed by the more traditional design of form.
JR: So it seems that, for you, computational design is not about eliminating authorship or subjectivity. You could perhaps even argue that computation has given you an expanded authorial freedom?
CB: Certainly. For example, the designs can exist in many different forms. It’s not just about producing an object, but about producing a field of outcomes. It’s really a celebration of simultaneity. From a simple set of relationships, I am able to make a huge variety of things, and I would argue that the outputs produced at any point in the process are equally valid. Sigma, for example, has produced computer animations, prints, and lost-wax casts, all of which I consider art, because they are all part of one process, one investigation—a kind of forensic aesthetic. A recent steel sculpture for the Salvatore Ferragamo Museum in Florence, Strange Attractor [2014], was made from the same abstract ingredients driving Sigma.
JR: It sounds as though your practice has evolved beyond any one particular medium—to the point where the programming and manipulation of data is your actual, primary design method.
CB: Yes. Data is a raw material for me, and I think it’s probably the most important new material we have today. And as Sigma’s forms show, data has specific qualities, like any other material. But it’s also crucial to understand that data by itself is meaningless. It’s only when you tie bits of information together in an algorithm or a sequence—when they begin to talk to one another—that meaning is produced. And that connection becomes the object of design. Data can connect simplistically, and the result will probably be crude. It can connect in sophisticated ways, and the result will be intelligent.
In the end, though, the systems themselves don’t matter as much as what they can produce. I’m most interested in the way my work can shift the viewer’s perception. With the Serpentine Pavilion, for example, or with the outputs of Sigma, no one needs to understand the algorithm to fully experience the potential of the form, or to sense that one is dealing with the relationship between shape, structure, and ornament in a new way.
JR: Then the question becomes: How do you translate the realm of data and computation into the realm of perception and experience? Scale seems to be particularly crucial: From Minimalism to contemporary architecture, it’s always been contentious—key to debates about spectacle, context, and public space. But the digital environment seems inherently scaleless. And the most common approaches to digital design today rely on more or less arbitrary translations of scale. One tendency is to create something that is purely digital—a complex form generated by an algorithm, say—and then pretend that it can be translated directly into physical form. In practice, this means that the design is produced at whatever scale is feasible for fabrication: small if it’s 3-D-printed, big if it’s CNC-milled. Or, many architects simply use computational technology to rationalize a pre-shape—whether it’s something they’ve sketched themselves or even borrowed from nature—so, again, there’s no meaningful feedback between the digital and the material. How do you address these problems of translation?
CB: Well, you’re right that it’s shortsighted to think that one can directly translate a design from the digital to the material, or to assume that any digital tool can just be applied to any given building at any size. In my work, there is a more complicated negotiation between these two environments. Going back to the Serpentine Pavilion, I began with this idea about intersecting lines. As that developed in a digital model, it wasn’t structure, it wasn’t space—it became a dimensionless set of interconnections. But I also knew the type of feeling that I wanted the visitor to have inside the building. That’s critical. In other words, I had a vision about the scale at which this intersecting geometry should be experienced. And that had certain implications: Say the square that’s generating the geometry becomes two hundred feet. Then each of the meshes becomes a certain size. And then you start to think about how the lines should read. Should they be thick and short? Should they be thin and run deep? If thin, then they should be steel, like blades cutting space. So once you choose scale, you’re into material and rhythm and texture. And then, finally, all this becomes fixed into a design.
But I should say that when it comes to perception, it’s not really the body—and certainly not questions of scale in terms of proportion—that I worry about anymore. That is quite classical. Instead, I would say the orientation and travel of the eye is one of my main concerns. And I find that there are ways to vary rhythms, to play with different axes of vision and shift the path of the gaze, that expand beyond a form’s direct relationship to the body, creating syncopated experiences that work simultaneously across multiple scales.
On my bridge in Coimbra, I produced what I think of as three speeds of perception. It’s a nine-hundred-foot-long bridge, but it’s designed with a major kink in the middle so that from either side you can see only the first half of it—the complete bridge disappears. The bridge deck is made of a structure of angled planks, and your eye can follow this into the distance until about a hundred feet into the bridge, where the planking becomes indistinct, and you lose the lines. Looking directly to either side, you see a system of colored glass that is calibrated to the length of your stride, which shifts approximately every three feet—with every step you take. So you got on the bridge thinking you were just going to walk right across, and within a few seconds you’ve paused—you’re noticing these different scales of looking, and your eye is brought from the length of the bridge to the breadth of your passage. Space intercepts you; you don’t simply pass right through it, as in a conventional bridge crossing. The whole idea of the bridge is to celebrate the river and enjoy the views of the old city.
All these different scales are rooted in the same computational system. But that’s completely secondary to the experiences that the system produced. By mediating between the virtual and the real, I can get people to experience space in a new way. And in one way or another, isn’t that what every artist is trying to do?
— Julian Rose talks with Cecil Balmond