IASS–nicholas goldsmith on shape finding vs. form finding
September 17, 2014 § 1 Comment
Yesterday’s keynote by Nicholas Goldsmith of FTL was a barn-burner; one of the best statements about design and engineering as intellectual discipline I’ve heard and a clear statement about what our tools can do versus why we use those tools. Goldsmith’s argument is that conceptual design—the part concerned with getting the overall scheme laid out and devising a strategy for solving the problem at hand spatially and geometrically—can fall into two categories. We sometimes try to find a ‘shape,’ that is, a formal geometry or algorithm that suits…something, either a program arrangement, a static principle, or our own aesthetically trained eyes. This accounts for a good bit of the work architects do, certainly, and it often accounts for engineering on relatively simple projects. The results are fairly straightforward statements, either sculptural, functional, or static. In contrast to this, though, is how Goldsmith defines form finding. This, he says, relies on Kant’s idea that form suggests that “every part owes its presence to the agency of other parts.” (I haven’t found this exact quote yet, but believe me, I’m looking). In other words, whereas shape implied nothing more than a convenient geometry, form suggests both an organization and a purpose. And this, of course, brings up D’Arcy Thompson, whose 1913 book on evolutionary biology, On Growth and Form, is required reading. The whole gist of Thompson’s argument was that organisms self-organize to produce useful structures based on simple algorithms—a radiolarian, for instance, that builds near-perfect geodesic spheres that minimize the amount of silica needed, or a nautilus shell that reproduces the same growth ring again and again, scaling up as it goes, to produce a shell from the simplest of genetically encoded instructions. The idea of teleonomy—production without foresight—is, according to Goldsmith, a model for form finding, in other words, the production of a global order from local actions.
So. Want to design a nifty-looking tent structure? Draw whatever you want—even build it—and you’ll get a reasonable shape. Want the most efficient tent structure? Let form-finding software crunch an algorithm to do with minimum surfaces and minimizing tensile stresses for a few hours, and see what patterns emerge. Or, interested in designing a tower that sheds wind vortices? Let the local actions—wind spilling off of a skyscraper façade—churn for a while, feeding back into software that changes the building shape and the cladding components based on finding the maximum performance. This evolutionary process will gradually eliminate all but the super-efficient options, leaving you with something more intelligently derived than just a shape. Form finding, in other words, implies development. Iterations that gradually eliminate bad ideas (“choice under stress,” in the words of Charles Eames) lead eventually to good ones (“how-it-should-be-ness,” again in Eames’ language). We’ve done this as designers for decades using physical models, but the discipline of evolutionary design has only been recognizably simulated with digital programs and feedback software like Grasshopper and Rhinoceros. And not only did Goldsmith name-check Kant, Thompson, and Eames in one lecture, he also cited my colleague Rob Whitehead’s work on Eero Saarinen, pointing out that the geometrically derived shell of Kresge Auditorium led to a predictably problematic structural and construction solution. His roof for TWA was similarly difficult to engineer and to build. It was only with Dulles, where feedback from engineers and contractors led to a shape that was based on production and performance, balanced against one another, where true form-finding happened. The discipline imposed by competing criteria force a truly responsive design process to filter out thousands of ideas that don’t work along one dimension or another, leading to sometimes surprising forms that do. Beauty, in this scenario, emerges from the embedded intelligence of the solutions—we recognize something engaging because of how fluently it balances the problems to hand. Goldsmith closed with Bucky Fuller’s quote, that he never thought about beauty until the end of the process, when if the solution wasn’t beautiful he knew it couldn’t be right. I think you could take that a step further and say that even beautiful solutions are sometimes not quite right, but as we get used to greater and greater complexity I think our standards of beauty must surely evolve along with the problems we’re solving. The idea that designs emerge out of any sort of disciplined approach–material, structural, even geometrical–and that the more dimensions the approach is responsible toward, the richer the design is likely to be dovetails nicely with everything I’ve found out about Chicago building, or about Nervi. Design is at its best when it’s an agile response to complex situations, and when it’s able to learn from interactions with the limits and suggestions those situations offer. Goldsmith’s work–which even he admits can veer from shape- to form-finding in the course of a project–summarizes that nicely and shows how even though the tools have changed, the responsibility of the designer to the facts on the ground and to a rigorous testing of and learning from ideas put out into the world remains unchanged. Great, great stuff. Especially after reminding ourselves in the morning that shapes in the right hands are also pretty engaging.