PRINT Summer 2011


IT’S BEEN A WHILE since architects talked about paintings. The long relationship between the two disciplines—one that ran from their premodern symbiosis through modern architecture’s entanglements with Purism, Constructivism, and De Stijl up to postmodernism’s fascination with Pop—has come undone. The legend of Le Corbusier painting in the morning and practicing architecture in the afternoon is no longer a disciplinary model but a historical curiosity. It is no coincidence that this parting of ways over the past two decades has occurred as architects have largely replaced physical drawing and modeling with computational representations, so that today younger architects can be found practicing architecture during the day and coding until the early morning.

For fifty years, research in architecture and related disciplines has pursued, and increasingly realized, a vision of a design process fully integrated through computation, in which a highly flexible model carrying all geometric, material, and performance data could be seamlessly translated into a material result. And though often burdened by architecture’s own versions of medium-specificity—fundamentalisms holding that the complexity of building can be reduced to function or structure or form or energy use—the main result of advances in computation-based architecture has been a proliferation of possibilities, an explosion in the range of what is buildable.

Ironically, then, the very computational tools that have severed architects’ most immediate connections to painting and drawing have also led to a formal and technical freedom for which the only significant historical parallel is, in fact, postwar abstract painting. Counterintuitive as it may seem, architecture today confronts problems similar to those faced by practitioners of postwar abstraction, from Frank Stella to Sol LeWitt, and the responses of these artists offer instructive templates. Oddly, postwar abstraction was, in its moment, almost entirely overlooked by architects, who—from the time of Robert Venturi’s Complexity and Contradiction in Architecture of 1966 through the Museum of Modern Art in New York’s 1988 “Deconstructivist Architecture” exhibition—instead focused either on the iconographic games of Pop or on reworking the legacy of the interwar avant-gardes.

As in postwar painting, the formal freedom that computation now offers architecture seems (at least to more thoughtful practitioners) to be of the “nothing left to lose” variety, prompting a motivational crisis to which self-imposed restraint is perhaps the only productive response. Methodologically, then, contemporary computational architecture reenacts the dilemmas of noncomposition and expression that were central to postwar art. As in postwar abstraction, contemporary architecture has come to be constituted by a great variety of approaches, each of which strives to be autonomous and internally systematic. Because projects can now be designed, at least in part, by code, by computational algorithm, the architect stands one level further removed from the realized building: no longer composing representations of reality, but manipulating computer scripts that generate representations of reality. The potential for systematic variation offered by scripts, and the sheer speed of computational representation, allows the exploration of potentially vast series of possible solutions for any one project. One result is that taste enters in a new way, as architects adjust limits and select versions, pruning and nurturing quasi-animate representations of projects, and becoming in some sense clients of their own semiautomatic design processes.

The dialectic of freedom and systemization manifests itself most clearly in the outer envelopes of contemporary architecture: the “skins” or “wrappers” to which the traditional term facade—with its suggestion of bounded frontality—no longer applies. Extrapolated both from Le Corbusier’s “free facade,” which was liberated from bearing loads, and from Robert Venturi’s “decorated shed,” which posited the facade as an iconographic appliqué to a container of generic space, today’s building skins are largely free to discover their own rationales. Formally, the systematic variations of contemporary architecture appear most often—and especially in skins—as fields: two- or three-dimensional arrays containing large numbers of elements coordinated by rules or patterns that are literally encoded within a computational model. It is this use of ordered yet irregular abstract fields that distinguishes contemporary architectural skins from their twentieth-century precursors. While highly variable, the fields are nonetheless systematic and therefore unlike the compositional and figural facades of Le Corbusier’s interwar villas. However, the great variation between—and, more important, within—these recent works also separates them from Ludwig Mies van der Rohe’s pursuit of a single universal and essentialist system. Though often used decoratively to wrap generic spaces, the abstract patterns of today’s skins set them apart from the historical and pop-cultural riffs of Venturi’s facades. Neither figural, essentialist, nor historicist, contemporary architecture has become a varied field of objects of varied fields, with each project establishing the rules of its own internal organization (here Peter Eisenman’s early houses—themselves in dialogue with Conceptual art—stand as important precursors).

Yet despite their parallels, strategies that were polemical and provocative in postwar art often seem diminished in computation-based architecture—accelerated and smoothed into a professional working method drained of intensity. If LeWitt famously claimed that for Conceptual practices, “the idea becomes a machine that makes the art,” then too often today it looks as if the (computational) machine becomes the idea that makes the building. Or perhaps it would be better to say that systems, series, and fields—those vaunted artistic methods of the postwar avant-gardes—have always posed risks of mindlessness. And this is where the example of postwar abstract painting is most valuable to contemporary architecture: as an intense, critical articulation of exactly this problematic. As Rosalind Krauss outlined in her 1971 Artforum article “Stella’s New Work and the Problem of Series,” the key issue is the inimical relationship between the material, phenomenal presence of a specific work and its simultaneous determination as the result of an abstract, conceptual process—what could be called the mind/body problem of serial work. In Krauss’s view: “The lesson that Stella seems to have drawn from his own work with serialization is that as the painting tends toward diagram, it enters the condition of the mathematical formula—both it and the developmental runs it is capable of generating are transparent to consciousness as an abstraction through time, in which both its presence and the viewer’s physical presence to it are no longer part of its meaning.” Where the reaction of Stella and his peers was to search for strategies that would resist, or at least complicate, this emptying out of the work’s presence, contemporary architecture has—with precious few exceptions—been propelled toward the ever more complete mathematization of its processes and their ever less examined realization. Increasingly, a building may be merely a representation of a computational model generated by an algorithm, with little regard for the bodies that inhabit it.

A tangle of issues surround this fundamental relationship of abstract conception and material realization—all of which should offer contemporary architecture possibilities for interrupting the overly smooth translation from the former to the latter. One problem that could be considered fruitfully is the relationship of field to shape—the former conceived of as infinitely extensible, the latter defined by the closed boundaries of a work. Taking an option not available to painting, contemporary architecture often chooses to defuse the potential tension between field and shape by wrapping all exterior faces, and perhaps even the roof, with a single continuous field. Lines, surfaces, and openings are demonstratively folded over volumetric edges, or the edges themselves are rounded out to create a single uninterrupted skin. Japanese architect Toyo Ito, for example, has produced a small portfolio of such buildings, including stores for Tod’s and Mikimoto in Tokyo, and the 2002 Serpentine Gallery Pavilion in London. While the problem of the edge must still be confronted at the base, and usually at the top, of the vertical skin, the main effort is directed toward a sense that the building is sheathed in a single uninterrupted field.

This approach evades the sort of critical engagement with shape that Michael Fried famously identified in the work of Stella and, to a lesser extent, Kenneth Noland and Jules Olitski. In the strongest of those examples, such as Stella’s complexly shaped canvases of the mid-1960s, the paintings explore the entanglement of what Fried termed “depicted shape” and “literal shape”: the outlines of visual elements within the pictorial field versus the actual shape of the support. An investigation into depicted shape and literal shape, then, means an inquiry into the work’s simultaneity as both an image field and an object. By defaulting to the use of continuous wrappers, contemporary architecture too often denies itself the possibility of such exploration. One notable exception is the Seattle Central Library by Rem Koolhaas’s Office for Metropolitan Architecture (OMA). Koolhaas has long had an interest in the interplay of field and shape, in the sense described by Fried, and the Seattle Library demonstrates a complex, purposefully unresolved relationship between its diamond-patterned diagrid skin and its multifaceted building volume.

A related and more pressing provocation, conspicuously dodged by much computational architecture, is us: the presence of people and our stubbornly corporeal bodies in confrontation with the abstract architectural field—what might be called contemporary architecture’s “allover” problem. Within a computational model there is effectively no marker of, and no limit to, scale, especially when the model is generated directly by code and not “by hand” (on a mouse, that is). When this limitless depth of scale is combined with parametric modeling, which tightly interlinks all elements, and with fabrication processes of mass customization, a new relationship of part and whole becomes possible—a highly intricate and totalizing organization that lacks any inherent sense of scale. Born in the scaleless world of pure geometry, computation-based projects are then faced with the troublesome problem of reintroducing the body and all of its architectural indexes: stairs, doors, ceiling heights, bathrooms, etc. (To a great extent this explains the tendency for computational generation to remain on the surface of buildings, where the constraints of human occupation can be minimized.) The most common response to what is, in fact, a deep antinomy in algorithmically generated architecture—it both is and is not for human inhabitation—has been to elide the difficulty by manipulating the systemic field so that necessary conditions “happen” to occur in the right places. Often the result is a painful sense of contrived coincidence: Take the voids at street level in Ito’s Mikimoto store, which are part of a purely formal field but are also all too conveniently in the right place, at the right scale, to serve as entries.

Again, contemporary architects would do well to consider the catalogue of art that began dealing with parallel problems decades ago. As a generation of critics have taken pains to point out, the works of artists such as Jackson Pollock, Barnett Newman, Stella, and LeWitt center on the mind/body problem presented by systems and series—and not as attempted solutions, but as engaged, and engaging, “worryings” of the problem. Indeed, one of the most pertinent examples for parametric architecture may be Pollock’s drip paintings, in which the range of scale among the marks is equally vast, and in which larger structures emerge out of the accumulation of smaller semi-independent ones. Importantly, and famously, the impact of Pollock’s allover fields revolves around their abandonment of traditional relational composition in favor of a paradoxical binding of scaleless, dematerialized optical experience to the scale of the body, its indexical markings, and the horizontality and materiality of that mark-making. Architects might also recall the demonstratively rough brushwork of Stella’s “Black Paintings,” which embeds scale, process, and materiality within the visual fields of the works. Or take the example of LeWitt’s wall drawings and paintings: As sets of instructions for the generation of form, these are the most literal precursors of today’s computational methods. It should be obvious that, though the works are algorithmic in conception, to automate the execution of one would be a coarse travesty, since it is only the friction of abstract thought and concrete act, rule and freedom, that gives LeWitt’s wall drawings and paintings their artistic heat. As computational architecture (hopefully) moves beyond its initial phase of unadulterated technophilia, we might imagine practices that work toward a parallel interweaving of the scaleless and the bodily, the abstract and the material.

One classic binary relationship that recent architecture has explored more deeply is that between the patterned visual field and its support—or, more plainly, the relationship of ornament and structure. Along with encouraging novelties of pure geometry, computational tools have also enabled the engineering analysis of many unusual structures, materials, and details that are entirely inaccessible to older methods of calculation “by hand”; and structural engineers, most prominently Cecil Balmond (formerly of Arup), have become true collaborators for many of the most celebrated architectural projects of the past two decades. They have done so by using computer-based analysis to realize schemes that are simply perverse from the viewpoint of traditional engineering. The vast, asymmetrical cantilever of OMA’s tower for the China Central Television headquarters in Beijing—to choose an iconic example—defies every standard of simplicity, stability, and efficiency. This flouting of good engineering sense is widespread in recent work and may be the strongest commonality among otherwise widely varying results. Here, in the most unexpected of places, we see an entire branch of the discipline embracing another of LeWitt’s famous maxims: “Irrational thoughts should be followed absolutely and logically.”

Recent work also overturns the traditional architectural hierarchy of structure (primary) and ornament (secondary or, for modernism, simply superfluous), which are now often combined in a single exterior layer driven by the needs of vision and support simultaneously. Though support clearly has different meanings in architecture and in painting, one can recognize a rough parallel between these recent architectural skins and the canvas manipulations of Piero Manzoni and Lucio Fontana, or Robert Morris’s felt pieces. In the best examples of each, the material means and visual ends are neither “ordered,” with one serving the other, nor “synchronized,” with the two massaged into convenient alignment, but are instead copresent, distinctly and with equal force. One recent architectural example of such copresence is the O-14 tower in Dubai by Reiser + Umemoto, which is wrapped by an undulating white skin perforated by a field of circular openings of varying diameter. The size and position of these openings reflect the influence of purely visual—and quasi–Op art—desires, but they also reflect structural demands that cause the openings to be organized roughly on a diagrid and with subtly greater solidity at the base than at the top. Interestingly, the firm thought that its initial attempts to design the pattern purely by algorithm appeared “too mechanical,” and the openings were ultimately placed “by hand” in order to achieve the right level of irregularity. Such nuanced variation from the pure abstract solution hints at the direction a mature computational architecture would take. As Jesse Reiser has said, “The computer allows us to work like painters.”

No one could say that, as a discipline, architecture has ignored the impact of advanced computational methods. If anything, over the past decade and more, the flood of books, journal issues, exhibitions, conferences, studio assignments, and project proposals has been enough to induce widespread fatigue (the evidence of which lies in the heap of overexposed terms—virtual, digital, algorithmic, parametric—that have already burned out). Yet these explorations have too often been solipsistic and isolated, both historically and aesthetically, as if the new technologies had generated their own contexts, ex nihilo, rather than having emerged within a dense preexisting cultural skein. At the same time, criticism of computational practices, especially within schools of architecture, has too often been simply reactionary, bluntly rejecting these developments via a more or less nostalgic humanism. All of which is understandable as an initial developmental phase, especially given the technical nature of computational work and the scale of disruption it has brought to the comparatively slow art of architecture. What is needed now, though, is the maturation of architectural practice and criticism beyond these isolated positions so that computational methods can be considered within, rather than simply against, historical and aesthetic contexts. Which is to say that it is time for architects to begin looking at paintings again.

Sean Keller is an assistant professor of architectural history at Illinois Institute of Technology, Chicago.