Research

Architecture recurrently draws inspiration from morphogenetic theories in contemporary scientific discourse. Responding to significant revolutionary studies in the natural sciences including the morphogenetic theories of Charles Darwin (On the Origin of Species, 1859), nineteenth century German architect Gottfried Semper drew a direct analogy between the evolutionary history of nature and that of architecture.  In Style in the Technical and Tectonic Arts; or, Practical Aesthetics (1860-63), Semper argued that art, and architecture in particular, is based on a few standard forms and types that derive from the most ancient traditions.  He posited textiles as the primeval art from which all other arts borrowed their types and symbols. Understanding of the “higher arts” is thus attained through analysis of the oldest and simplest works of technical art –including textiles—where the preconditions of form are most apparent and comprehensible.  More than a century later, British architect John Frazer’s seminal text, An Evolutionary Architecture (1995) paralleled the continuing search for a unifying theory of morphogenesis in the biological and physical sciences.  Responding to the newer sciences of cybernetics, artificial intelligence, complexity, and chaos, Frazer proposed breeding as opposed to designing new building forms.  “Genetically representing” architecture in a DNA like code-script, Frazer considered architecture as a form of artificial life, subject to developmental and evolutionary processes in response to user and environment. Like Semper, Frazer focused on prototypical forms.  He used computer models to simulate many generations of populations of simple forms undergoing replication, mutation, and selection based on their performance in changing environments. His long term goal was to literally incorporate the processes and materials of building into the model so that the emergent[1] structures would be self-constructing. This dissertation will address the shift in representation inherent in the evolutionary architectural model; one that alters the process of design—privileging starting conditions over process or final product—and challenges the traditional role of the architect.

 

Marco Frascari argues that the objects of traditional analogical design procedures—including process drawings, sketches, notes, and marginalia—are “genetic drawings.”[2]   They activate and trace both mental structures and physical actions and as such, are both catalysts for and reflections of a vision of the future construction.   According to Frazer’s paradigm, the architect is primarily responsible for the explicit definition of what he terms a “genetic representation.” This representation specifies the genetic code-script of the virtual genes, the rules by which the code evolves, and the relationship (mapping) between the virtual genes and the virtual appearances, properties and behaviours they will generate. Almost everything else—tabulating replications, assigning fitness to new variants, and selecting which genes will be passed on to the next generation is the task of a specialized class of search algorithms[3] known as “genetic algorithms”.

This dissertation will develop the notion of algorithmic generative systems in architecture; systems that produce a wide variety of patterns, are integrated with their building materials and construction systems, interact with and respond to both designer and user, and adapt over time. Parallel theoretical and physical investigations will examine the cognitive and corporeal operations and outcomes by which “genetic representations” are generated, grown and delivered. Research will survey the scientific theories and technological sources that influenced Frazer and his contemporaries working at the nexus of computing and architecture between 1970 and 1990, and the theories and technologies influencing the use of generative systems to design architecture today. This dissertation will examine the expression of these concepts in a number of case studies: Frazer’s “Universal Constructor” and “Universal Interactor” and contemporary projects including the Beijing Olympic National Aquatics Center , known as the “Water Cube” ([H2O]³), by PTW Architects and Ove Arup.

Synthesizing Semper’s approach—textile analysis as a means of elucidating the fundamental principles of architecture—with Frazer’s desire to integrate simulation with construction, a series of physical models of generative systems will be explored using computer aided knitting machines as output/input/feedback devices. Knitting is particularly well-suited to the study of generative systems because unlike most other fabric structures, a knit fabric is capable of being three-dimensionally shaped during computer aided manufacturing without the need for post-production cutting or sewing. Knitting generates a fabric by looping, one loop at a time; a type of cellular construction created by the aggregation of units. Simple variations in loop arrangement –such as whether the loop is worked or not, and whether it is drawn to the right or the left—are capable of being ordered in an infinite number of patterns. Knitting patterns operate like algorithms.[4]  The repetition of specific sequences of procedures produces a countless variety of textures, surfaces and form. Knitting is a self-organizing process; structural and material properties are emergent behaviours arising from the act of fabrication.

 

There is a clear contrast between the limited results which architects have so far obtained with genetic algorithms and those achieved by biological evolution. Manuel DeLanda contends that the productive use of genetic algorithms implies the deployment of three forms of philosophical thinking (populational, intensive, and topological thinking). Frascari asserts that the “…textile is in itself a construction of critical knowledge. This knowledge is based on the twofold being of any product generated by a technology, the processes of construction and construing.”[5]   Recent studies in neuroscience suggest that emotions and the bodily manipulation of external vehicles are also constitutive of cognitive processes.[6] The development of an effective set of conceptual and philosophical tools with which to think critically about the genesis of form and genetic representation will be a significant measure of the research.   Ultimately, the thesis will seek to draw a clear picture of genetic representation; how it is changing the way we construe and construct contemporary architecture.

[1] Manuel DeLanda describes the modern concept of “emergence” as a phenomenon in which novel properties and capacities arise in a causal relationship.  (Manuel DeLanda, Philosophy and Simulation; the Emergence of Synthetic Reason (London: Continuum, 2011)3.

[2] Marco Frascari, Eleven Exercises in the Art of Architectural Drawing: Slow Food for the Architect’s Imagination (Abingdon, England; New York: Routledge, 201) 90.

[3] An algorithm is an explicitly defined set of procedures.  In the context of Frazer’s “genetic algorithm”, the term refers to the text of some computer code.

[4] It is interesting to note that the earliest history of computers and algorithms is marked by a kind of textile thinking. Charles Babbage, who first theorized the computer, intended to use the system of punched cards (invented by Joseph Marie Jacquard to control the execution of complex patterns in mechanized knitting and weaving) to control his Machine Analytique. The program Ada King, Countess of Lovelace, wrote for Babbage’s Machine is now recognized as the first algorithm.

[5] Frascari, Monsters of Architecture (Savage, Maryland: Rowman & Littlefield Publishers, 1991) 118.

[6] See Antonio Damasio’s The Feeling of What Happens: Body and Emotion in the Making of Consciousness (New York: Harcourt Brace, 1999) and Richard Menary, ed. The Extended Mind (Cambridge: The MIT Press, 2010).