The idea of evolution in digital architecture: Toward united ontologies?

Computational and natural as separate ontologies

The protagonist of evolutionary understanding in digital architecture was John Frazer. Utilising the Neo-Darwinian evolutionary model, Frazer attempted to naturalise architectural design by transferring information from genetics and evolution mechanisms. Frazer ³⁰ utilised dynamic, self-renewing and responsive characteristics of evolutionary nature in the architectural design process via 3D evolutionary cellular automation. This system tested emergent forms through the accelerated virtual evolutionary model shaped by the internal dynamics of genetics and external environmental factors. When the complexity of the design reached a certain level in the evolutionary process, the architectural forms could organise themselves and produce more complex copies and structures, just like a natural evolution. Frazer understood computational ontology as an additional dimension that allowed designers to generate forms through the evolutionary process. Computation was not a replacement for the real world.

In Frazer’s understanding, evolutionary processes were considered the essential element even more important than the design objects’ performative, functional and aesthetical qualities. Thus, the evolutionary design process represented a significant paradigm shift regarding the architect’s authority since the design was left to its own development in the creative virtual evolutionary environment. The emerging forms were unpredictable since the characteristics of randomness were introduced into the bottom-up evolutionary process through variation mechanisms and environmental factors. Therefore, the natural model manifested itself by directing the development of the architectural form as abstracted representations of evolutionary mechanisms and contributing to the selection process by providing feedback in the form of digital data collected from environmental conditions.

Greg Lynn accepted endosymbiotic theory as the best evolutionary model for the animation-based understanding of form by criticising the Neo-Darwinian evolutionary model for its dependence on random mutation, external natural selection, gradualism and reductionism.^31,32 Against the population and gene-based evolutionary model of Neo-Darwinism, the endosymbiotic theory highlighted metabolism and community-based model that explains the bottom-up evolutionary process where single organisms merge with neighbours to form complexity. ³³ Such a perspective on the evolutionary process based on symbiosis let Lynn ³⁴ recognises architectural geometry as an evolutionary space where continuous differential transformations occur. Therefore, architectural form represented dynamic, smooth and mutable singularity depending on the changing reciprocal relationship between continuous iterative differentiation and external evolutionary dynamics. ³⁵ The reflections of this understanding were evident in the Embryological Houses project, where the constant mutated models were proposed. ³⁶

Lynn promoted the importance of process as endless variations of the form. All final forms were considered perfect in their unique contexts since the logic of continuous change depending on external forces was proposed to justify the design solutions. As a direct impact of evolutionary process-based perspective, architectural form was recognised as momentary responsiveness rather than a static entity disconnected from its environmental forces. This realisation demonstrated that the evolvability of form was the substantial reason for the complexity of the system. Evolutionary randomness did not play a critical role. Like Frazer, the evolutionary logic contemplated within the boundaries of the computational ontology and convergence with the outside world was not investigated. Symbiosis as an evolutionary process only provided abstract principles and design logic to be used in the transformation of the architectural form.

Karl Chu proposed genetic architecture by fundamentally linking the plane of immanence, endosymbiotic theory and cellular automation. The plane of immanence was described as the foundation of philosophy that produces infinitive concepts, movements and variation. ³⁷ Accepting the plane of immanence as metaphysical construction that embodies possible dynamic structure interactions composed by monads, Chu ³⁸ introduced the idea of a ‘cone of immanenscendence’ as an ontological rendering that launches perpetual worlds. Following this theoretical framework, Chu ³⁹ proposed genetic architecture to describe a novel post-human agenda that phases out anthropocentric architecture views. Chu categorised two diverse perspectives for the embodiment of computation in architectural design: morphodynamic and morphogenetic. In the morphodynamic approach, form evolves according to external dynamics, whereas in the morphogenetic one, form emerges through the internal logic of genetics, such as self-replication and reproduction.

Karl Chu criticised the morphodynamic approach as it failed to recognise objects with internal structures and works with pre-defined quantitative values, preventing randomness. The openness and random qualities of the computational system were essential since the natural events cannot be reduced into structured parameters. ⁴⁰ By understanding nature as a system based on monads that carry irreducible information and act as self-replicating units, Chu conceptualised architectural design as a co-evolutionary process evolving from simple assemblages to complex self-organising systems. Nevertheless, Chu’s design theory was limited to digital design experiments, failing to show the physical manifestation of such a theoretical model. Also, this bottom-up evolution from simple parts to complex arrangements through digital computation procedures includes only abstract principles derived from the working logic of symbiosis in nature.

The Emergent Technologies and Design Group formed by Michael Weinstock, Achim Menges and Michael Hensel further developed morphogenetic design theories. They proposed a design methodology that employed evolutionary and morphogenetic processes for higher performance, functionality and environmental sensitivity.^41,42 Random evolutionary variations were used to bring stochastic quality and diversity to the design process, boosting the system’s robustness and complexity. Building an analogy with evolutionary process in nature, fitness criteria were applied to evaluate candidate designs regarding their performance for finding optimum solutions. Also, material behaviour was utilised as a generative and selective design instrument by creating a complex multi-performative system informed by internal and environmental factors.

In this morphogenetic design model, architectural form resulted from the process instructed by population-based evolution, self-organisation procedures and material agency. The proposed model differed from Frazer’s regarding the role of digital computation and active participation of material behaviour. Randomness was introduced to the model through variation mechanisms to expand search space and the data obtained from the material agency. Therefore, evolutionary processes were not used to generate mere variations of the architectural objects but to create architectural models where the formal, structural and material qualities are interlaced, like the organism in nature.

Architect Evan Douglis combined Neo-Darwinian gradual evolution and William Bateson ⁴³ ’s discontinuous change through computational design process. The design model was initiated with a generative computational unit capable of creating limitless variations through repetition depending on dynamic environmental forces. ⁴⁴ Douglis constructed an evolutionary design process, both directive and progressive but not deterministic since controlled accidents or mutations were introduced to the system for unpredictable expressions of structural surfaces. ⁴⁵ At every stage of the process, designs were subjected to the fitness test, measuring the performances depending on the potential to be combined with other units and create three-dimensional complexity. ⁴⁶

Douglis instrumentalised the mutation mechanism in the design process as an explorative tool that expands the search space and introduces random qualities to the system. The evolutionary process was conceptualised as the creation of endless variations of surface structures through bottom-up procedures depending on the co-evolution of units, resembling the form generation in cellular automata. Still, the model differed from digital cellular automation since the material behaviour of the units also played a crucial role in the development of the form through analogue techniques. The autogenic structures were evolutionary form-finding experiments searching for indeterminacy, emergence and complexity through both digital and material computation.

The architect Alberto Estévez proposed cybernetic-digital as the logic of virtual DNA to link the binary system of the computation with the organisation of natural genetics. ⁴⁷ Such a model focused on creative computational design procedures to explore the potentials of the genotype-phenotype relationship of the living world. Estévez advocated going beyond the mere imitation of static natural forms by concentrating on computational models of biological evolutionary processes, which simulated growth, self-organisation and morphogenesis. However, Estévez did not provide further critical thinking to explain the creative and intellectual role of evolutionary process, except for an analogy between the emergence in nature and computational design.

Computational and natural as united ontologies

Nature is an open system evolving under the influence of irreducible divergent parameters. On the contrary, computational methods tend to be logical, prescriptive and autonomous. While the analogue computation graphs continuity as an image of the real through semantics, the digital computation utilises discrete elements and discontinuous scales to represent the real. ⁴⁸ Whether analogue or digital, the computational approaches regarding the representation of evolutionary process can only provide divergent degrees of abstracted versions of reality. ⁴⁹ One must overcome this essential struggle to create evolutionary design procedures uniting the computational and natural ontologies. These two can only merge through the epistemological framework of the designers, which is subjected to several theoretical approaches.

The architect Dennis Dollens proposed one of the earliest examples of united ontology. Dollen initially accepted Louis Henry Sullivan’s drawings as seeds providing design instruction and combined it with the computational analysis of plant growth in the design process.^50,51 Following this reasoning, Dollens conceptualised another framework by understanding monads as computational instruments carrying evolutionary potentials in the plane of immanence, and utilised extended phenotype theory to unify this system with the physical world.^52,53 In further investigations, Dollens ⁵⁴ also questioned the idea of thinking buildings as a quest for an intelligent built environment. Employing the autopoiesis theory as a platform that organises the information transfer from nature, ⁵⁵ Dollens also used the extended mind hypothesis as environmental cognitive theory linking nature, computation and design objects as interconnected existence.^56,57

Dollen’s main argument was the hybridised bio-digital ecology where the buildings were considered dynamic and intelligent systems continuously evolving with nature. From the computational representation of evolutionary processes to its utilisation by architects and digital design fabrication, Dollens envisaged the design process as part of the same ontology united with nature. Beyond the different utilisations of the evolutionary process in form generation, Dollens directly focused on unifying architecture and nature through computation, cognition and evolution theories. Nevertheless, extended phenotype and extended mind hypotheses, the mediator between ontologies, are still debated regarding the ontological problem, and the computation is only a powerful tool to represent nature thanks to its processing power.

The architect Alberto Estévez ⁵⁸ also proposed ecological-environmental design aiming to modify living organisms’ DNA through the logic of computation to design habitable places. Future biological buildings are expected to have programmable living materials, elements and spaces. This vision has remained a utopian imagination that has shown no remarkable progress. Estévez’s projects failed to achieve the central evolutionary promise of designing an architectural space by manipulating living genetic material. How the binary logic of computation converges with the working mechanisms of genetic material and opens the possibilities of robotised control on the growth of the living building are not adequately explained.

To overcome abstracted and prescriptive evolutionary strategies, Alisa Andrasek adopted the triadic logic by C. S. Pierce and utilised the brute force of discrete computation to obtain massive data from the core fabric of material agency. ⁵⁹ Unlike binary logic, triadic logic introduces a three-valued system; true, false and indeterminate. ⁶⁰ For Andrasek, ⁶¹ the third value of triadic logic and computational data obtained from material agency brings randomness into the design, providing an opportunity for hybridised ecology by interlacing living and non-living intelligence. Understanding matter as information, Andrasek utilised discrete computation to propose an architectural complexity and aesthetics beyond human cognition.

Although many studies examine the evolution of material agency and microstructures, there has yet to be a consensus on a general framework explaining the connection between micro and macro evolutionary processes. ⁶² Andrasek attempted to unite these two scales by using pattern-recognised information obtained from the material behaviour in the design, introducing evolutionary process from the micro-scale. Nevertheless, when the design is manufactured, it fails to respond to the natural environment since the collected computational data is about the context of micro-level behaviour, not the organism and system level. Also, the pattern-recognised information provides a data-rich representation of matter beyond the centralised evolutionary process. However, it is still a degree of abstraction and simplification, not the material itself.

Architect Jenny E. Sabin and biologist Peter L. Jones spanned the boundaries between biology and architecture through computation, simulation and digital fabrication technologies. ⁶³ They analysed the emergence of natural complexity at different scales, such as cell structures, morphological/molecular changes, self-organisation and abstracted these models into dynamic computational design models. Understanding biological events as information structures and processes, Sabin built an analogy with Conrad Waddington’s epigenetic landscape representing the embryonic development of the cells. Therefore, Sabin ⁶⁴ proposed a matrix architecture, where the algorithm illustrates the design of dynamic forces through a diagram coding developmental landscapes' potential networks.

The most promising argument proposed by Sabin was understanding design as a continuous evolutionary process of forces instead of objects and their relations. Therefore, the object was only a secondary quality of the evolutionary process both in nature and computational design. Sabin believed that computation could be used as a mediator to bring the processual operation of nature into the architecture, allowing designers to embody natural ontology. Antoine Picon ⁶⁵ defined such an approach as reinventing nature. However, the utilisation of digital computation is essentially a reduction of the qualities of natural process into information bits.

There is also an emergent research area where the Physarum Polycephalum’s collective foraging, extension and adaptation behaviour has been observed and applied as a problem-solving method. ⁶⁶ Physarum Polycephalum is a large, single-celled, multi-headed slime mould composed of tube-like structures creating an optimised network connecting the food sources. ⁶⁷ The organisms’ creativity stems from an embodied intelligence that depends on material interaction and its relationship with the environment. ⁶⁸ Physarum was used for several purposes in architecture, from road network optimisation to 3D form-finding experiments.^69,70 It was also considered a bio-digital aesthetic principle that defines biotechnological wholeness and ecological intelligence in which non-human actors were co-designers. ⁷¹

After successive stages of the natural selection process, Physarum has gained distinctive features to optimise dynamic morphology on the planar, topological and volumetric plane. Indeed, the organism is not trying to solve any computational problem, just struggling to survive. The evolutionary perspective highlights how organisms' will to survive can be utilised as co-designer in architecture. Physarum computation is a cutting-edge understanding in digital architecture in which evolutionary and morphological mechanisms are no longer centralised and universalised, but their complex interactions are observed in behaviour, bias and material sense of the organism. Nevertheless, once the behaviour of Physarum is described as an algorithm, it becomes part of the computational ontology and represents only pattern-recognised information.

By integrating the living material into designed parts, neoplasmatic design proposed a semi-living understanding of the ecology. ⁷² Such a theory aimed to create bio-composite ecology by investigating and integrating the botanical matter, animal flesh and living conditions into the current built environment.^73,74 Following a similar understanding, bioreceptive design theory was proposed by Richard Beckett and Marcos Cruz. Initially, evolutionary and morphological growth processes were scanned and scripted to construct essential frontiers of the project. Then, the computational simulation tools were utilised to anticipate the growth of living material on the previously designed form. ⁷⁵ Bioscaffolds were used to integrate the geometries of living organisms into the design process. ⁷⁶

In bio-integrated design, the computation generated form through the evolutionary models and illustrated the growth and behaviour of living conditions on the same structure. Absolute design control was impossible since living creatures have their own autonomy, directly bringing the evolutionary process from natural ontology to architectural design without any abstraction. Still, computation only represented an abstract representation of the living world, meaning that designers cannot fully anticipate future behaviours. For this reason, the natural and artificial could only merge at the conceptual level, thanks to the computational bridge. Also, the central argument of unified ecologies composed of natural and artificial elements in continuous evolution was not achieved since bio-composites were controlled within closed habitats, isolating the system from the rest of the environmental conditions.

In the more recent united ontology attempt, Evan Douglis considered building parts as intelligent elements responding to the ever-changing conditions of planet earth. Blurring the boundaries between the living and non-living, Douglis ⁷⁷ considered the responsive and dynamic material hybrids the future of the built environment in the light of the Anthropocene era. In this cyber-physical ecology, the evolutionary process was expected to occur in real-time depending on the relationship between inherent material behaviour and environmental factors, introducing divergent degrees of randomness beyond human cognition. Therefore, the design resulted from the collaborative intelligence of humans, machines, materials and the environment. Since the building is part of the evolutionary process, abstracted models were not planned to be used in the design process. However, Douglis did not explain in depth the collective working procedures and feedback mechanisms of this fusion between computation and nature for the built environment.