I gave a talk on images of abstraction and abstractions of images in the two-day symposium “Computational Design: Practices, Histories, Infrastructures”, organized by Daniel Cardoso Llach at Carnegie Mellon University, 7-8 October 2017. The event brought together historians, designers, and media artists in stimulating conversation about cultures, technics, and prospects of design computing.
Abstract: In 1961 architect and mathematician Christopher Alexander presented an audience of building scientists with a figure consisting of points and lines. This figure was not a geometric shape, but a mathematical entity that Alexander identified as a “linear graph.” Alexander enlisted the graph to “picture” the abstract structure that he saw as undergirding a “design problem” — a set of requirements to be met by a designer. He then presented a method for transforming what was a disordered entanglement of requirements into a neatly ordered “tree.” By the mid-1960s the tree would come into the parlance of architectural research as stand-in for the hierarchical nature of design processes and their physical outcomes, only to be soon dethroned by Alexander himself. Other graph manifestations —“simplices,” “semi-lattices,” “cascades,” “networks” — figured as correctives, signposting new eras in Alexander’s architecture-theoretical activity. In this talk, I track pictures of graphs in Alexander’s body of work from 1958 to 1974. Taking these pictures as instantiations of a mathematical entity with broader symbolic and operational attributes, I discuss the changing status of graph theory in the 1960s. Ultimately, I interrogate the forces that brought graphs into Alexander’s work as a way of concretely practicing structural abstraction — the lingua franca of architectural and mathematical modernism in the Postwar.