CITA PhD thesis - Jacob Riiber 2013 - Generative Processes in Architectural Design

Page 1

Generative Processes in Architectural Design Jacob Riiber

The Royal Danish Academy of Fine Arts, Schools of Architecture, Design and Conservation School of Architecture

Generative Processes in Architectural Design PhD Thesis by Jacob Riiber


Generative Processes in Architectural Design PhD Thesis by Jacob Riiber


Generative Processes in Architectural Design PhD Thesis by Jacob Riiber Supervisors: Professor Mette Ramsgard Thomsen Associate Professor Peter Bertram Published by: The Royal Danish Academy of Fine Arts, Schools of Architecture, Design and Conservation, Philip de Langes Allé 10 DK - 1435 Copenhagen K Print: The Royal Danish Academy of Fine Arts, Schools of Architecture, Design and Conservation, ISBN: Copyright © 2011 Jacob Riiber and The Royal Danish Academy of Fine Arts, Schools of Architecture, Design and Conservation



ABSTRACT


ABSTRACT

Observed in the context of a computational and generative approach to architectural

design the intention is to explore a relation between technology and difference emerg-

ing within this field. This endeavour is staged as an investigation of a productive conflict contained within a development of the concept of reflexivity. The conflict in question appears within a seemingly oppositional schema, defined firstly by concepts of plan-

ning, control and hierarchy, and secondly by notions of self-organisation, non-linearity and autonomous agency. By introducing the mechanism of reflexivity, this apparent

binary model is proposed to become directed towards something more ambiguous and complex, than what its simple construction might at first convey. Here reflexivity

is identified as a dynamic component articulating a process of mixing a necessity for directionality, overview and target with the unruly processes and agencies mediated by

generative design procedures. This conflicting structure defines the problem domain of the thesis, and the mechanism of reflexivity the core issue to be addressed.

Directed towards an understanding of this productive conflict the inquiry is conceptually approached as an effort to describe a suitable framework for a digital design

practice, capable of navigating the space defined by the problem domain. The concept of a framework is to be understood as a productive open-ended field – reminiscent of

a map – supporting, describing and enclosing a set of interrelated points of orienta-

tion and areas of movement. These are thought of as a means for steering a practice of design. Thus, the framework contains elements pertaining to methodology, technology and ideology without being either. The construction of the framework unfolds as an investigative, experimental and inventive practice by means of parallel movements

between discourse and making; accordingly, producing elements that are material, technical and conceptual.

The goal is to contribute to a tendency in contemporary architecture, which criti-

cally and experimentally explores potentials associated with the use of computational technology in architectural design. In this endeavour the project results in a series of

proposals for how such a practice might unfold, given a generative approach to design. Not as a methodology but as points of productive orientation, and as a supplement rather than substitute of existing practice.

5


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

Danish Version:

Set i perspektivet af en digital og generativ tilgang til arkitektonisk formgivning, er

afhandlingens hensigt at undersøge en relation mellem teknologi og forskel fremkommende

inden for dette felt. Dette forehavende udfoldes gennem udforskningen af en produktiv konflikt indeholdt i en bearbejdning af begrebet refleksivitet. Denne konflikt fremkommer

gennem en øjensynligt modsatrettet skematik defineret gennem først planlægning, kontrol

og hierarki, og dernæst af begreber som selv-organisering, ikke-linearitet og en kapacitet af selvstændig virkekraft (autonomous agency). Ved at introducere refleksiviteten som en

operativ mekanisme foreslås det, at denne umiddelbare binære model kan rettes mod noget

langt mere flertydigt og komplekst, end dens simple konstruktion på overfladen åbenbarer. Her identificeres refleksiviteten som en dynamisk komponent, der evner at artikulere en proces af sammenblanding mellem først retningen, overblikkets og målets nødvendighed

og dernæst de uregerlige processer og fremtrædener, der kan formidles gennem generative

fremgangsmåder. Denne modsætningsfyldte struktur definerer afhandlingens problemfelt og refleksivitets-begrebet definerer dens kerne-problematik.

Rettet mod en forståelse af denne produktive modsætning, er undersøgelsen idémæssigt

udformet som en bestræbelse på at beskrive en passende ramme for en formgivningspraksis, der evner at navigere i afhandlingens problemfelt. Begrebet ramme skal forstås som et åbent felt – med lighed til kortets begreb – der understøtter, beskriver og afgrænser et sæt

af indbyrdes forbundne orienteringspunkter og bevægelsesfelter. Disse elementer tænkes som midler med hvilke en formgivningspraksis kan styres. Således indeholder rammen

elementer, der relaterer til metode, teknologi og ideologi uden dog at være nogen af delene. Konstruktionen af rammen udfoldes gennem både diskursive og fremstillende bevægelser – som en undersøgende, eksperimentel og opfindende praksis. Således produceres elementer, der er både begrebslige, materielle og tekniske.

Målet er at bidrage til en tendens i samtidsarkitekturen, der kritisk og eksperimentelt udforsker potentialer forbundet til brugen af informationsteknologi i den arkitektoniske

formgivning. Med denne hensigt resulterer projektet i en serie forslag, til hvorledes en sådan praksis kan udfoldes i lyset af en generativ formgivningstilgang. Ikke som metode, men som produktive orienteringspunkter og som et supplement, snarere end en erstatning, til eksisterende praksis.

6


ABSTRACT

7


ACKNOWLEDGEMENTS


ACKNOWLEDGEMENTS

The completion of this thesis would not have happened without the help of the countless people that have generously guided and helped me these past years. Their positive attitude and support have made this vast task possible.

Firstly, I would like to thank my principle supervisor Mette Ramsgaard Thomsen for her

enormous commitment and level of ambition in tutoring me throughout the development of this thesis. Mette should also be acknowledged for her role in establishing the inspiring research environment of CITA, without which this project could never have happened.

I also want to offer my appreciation to Peter Bertram for acting as secondary supervisor and for substituting as principle supervisor during the finalizing of the project, during Mette’s

maternity leave. Peter’s sense of overview, productive criticality and calmness has been indispensable.

In the early stages of the project a number of inspirational conversations and brief collaborations with Jenny Sabin helped me greatly along the way. I regret that this time was so limited.

I would also like to thank Phil Ayres for sharing his knowledge of the field of cybernetics and for offering comments while writing this final document.

A special thank you should be addressed to Martin Tamke, without whom I would not have

entered the field of architectural research and for engaging me in the Lamella Flock project, which greatly inspired the present work.

I would also like to thank the entire CITA ‘team’ (past and present): Karin Bech, Paul Nicolas, Stig Nielsen, Anders Deleuran, Odilo Schoch, Annica Ekdahl, Johannes Beck, Andrius

Vilčinskas, Kristjana Sigurdardottir, Henrik Leander Evers, Holly Gibbons, Roxana Aron, Andrea Foged Trieb, Aron Fidjeland, Martin Toyberg Frantzen, Morten Bülow, Morten

Winther, Erin Towsley. In this context I will greatly miss the company of my fellow CITA

PhD students (also past and present): Brady Peters, Tore Banke, Aurélie Mossé, Anders Hermund, Norbert Palz, Dave Stasiuk, Vicki Thake.

I am of course also grateful to The Royal Academy of Fine Arts, School of Architecture, especially the support of Jørgen Hauberg, Lise Steiness and Vibber Hermansen – as well as the many students with which I have become acquainted.

During the production of the ‘Scatterings’ project I was generously supported by Malene Hvidt and Hans Bærholm – much more than could have been expected.

Beyond the context of the thesis period itself I would also like to acknowledge the significant impact of Cort Ross Dinesen who, amongst others, tutored me during my master thesis.

Lastly, a very large thank you to my supportive and patient wife Nanna, and to our sons Vilfred and Viggo

9


CONTENTS


CONTENTS

Preliminaries

1| Introduction

18

Towards the Procedural and Generative

20

Outline

24

Generative Design

22

Contextual Framing 24 Tendencies of a Critical Digital Practice

24

26

Generative Architecture

Material Agency

25

Procedural Complexity

26

Cultural Embedment

27

Constructing the Inquiry 28

Hierarchy

29

Problem Domain: The Concept of Reflexivity

31

Self-Organisation

Loss of Hierarchy

Aim

30 31 33

Contribution: A Framework for Design 34

Limits 34

Overall Construction of the Framework

Discovery as Method

35

Constituents and Target of the Framework

36

Questions

36 37

Thesis Structure 38

2| Methodology

44

Structure of the Methodology 46 Operational Framing

46

The Framework

49

Productive Conflict

49

Directional Structure: Points of Orientation

11

51


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

The Concept of Practice 52 Practice of Writing

54

56

Practice of Making

55

Experiments and Probes

Outline 58 Part 1

3| Technology beyond Optimization

62

Design, Technology and Optimisation 63 Preliminary Remarks on Optimisation and Technology

66

Two Models of Society 67

Societal Model I: Homeostat 68 Intermission: Criticality, Resistance and Alternative

74

An approach to Technology

79

4| A Manifold Digital

82

Societal Model II: Algorithmic 75

(The Question of) Unity Tendencies within the Field

85

Material as Agency

96

The Generative as a Design Approach

Information Technology and the Total Processes of Architecture

90

92 99

Cultural Reflexivity

104

107

Change From an Outside

104

Change From Within

Aligning the framework 108

5| A Cybernetic Approach

112

Compatibility: Design as Cybernetic System

116

Design Methods

12

116


CONTENTS

(In)compatibility 118 An Understanding of Cybernetics

119

120

Mechanisms of Adaption and Enabling Control

Second Order Systems and the Importance of the Observer

A Practice of Adaptive Modelling

Design and Cybernetics

Historical Parallels between Design and Cybernetics

Complexity and Solution

Design as Cybernetic Conversation

Synthesising with Cybernetics I

119 121 122 122 123 124 126

Contributions: A condition, a space and a mechanism

127

Non-representational technology

130

A Decentred Agency within an Aesthetically Potent Environment Observers, Reflexivity and Autopoiesis

127 134

Synthesising with Cybernetics II

140

Part 2

Experiment 1 | Lamella Flock

146

Preliminaries 149 Project Execution 152 Early Probes

152

The Program

154

Behaviour and Method

The System: Principles, Structure and Behaviour

Implementing Production Constraints

Designing Material Evidence

154 156 159 159

Reflections 160

6| Representation and Agency

166

Representation 168 From Being to Becoming 172 13


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

Morphogenesis 173

Multiplicities 174 Intensive and Extensive

177

Journey 180

Reflexivity and Cartography 181

Cartography 182 Field conditions

185

Experiment 2 | Porous Ascend

192

Representation and the Mechanism of Agency 189

Preliminaries 193 Research/Experiment Context

195

The Pattern

196

System 196 The Tool

198

The Making 200 Preparing the data

200

Unfolding for production and building

201

Production 204

Building 204

Findings and Challenges 206

6| Environment and Interface

210

Environment without Exteriority 211

Context 212

Environment 214

Environment as Double Articulation, Algorithmic Structure and Virtual Being Double Articulation

216

217

Virtual 219 Abstract Machines

220

Open and Systematic Montage

223

14


CONTENTS

Environment and the Becoming of Form

225

Experiment 3 | Scatterings

230

Project overview

233

Agent Space

235

237

The Design Space as Environment 226

Digital Design Spaces 234

Forming Connections

System Usage

235

The Makeup of Connections

238

Simulation Space

238

Component and Assemblage 240 Appearance

240

Fabrication

243

Materiality

243

Assemblage 245

Findings and Challenges 245 Ending

8 | Framework

252

Summarising the Problem Domain

253

A Practice Aligned to the Framework

258

Closing Remarks

262

Solidifications: Propositions and Generalisations 254 Currents: A Territory of Further Investigation 258 Field

259

Bibliography

264

List of Illustrations

270

15


PRELIMINARIES



1| INTRODUCTION


INTRODUCTION

With the present work I aim to contribute to a tendency in contemporary architectural practice, which critically explores potentials associated with the use of compu-

tational technology in design. The discussions and explorations presented here are a continuation of a long-lasting and personal interest in this field; an interest that I have

been able to maintain because of the inspiring manner by which a digitally produced architecture – continuously and unpredictably – develops and matures. For me this

process has been characterised by a movement departing from speculative experiments of form and aesthetics, towards also including a distinct coupling to our concrete and complex surrounding reality. A joining that I observe, amongst others, to be expressed by the way material, production and environment can now be mediated by

our architectural models, and consequently may enter dynamically into a total process

of design. I hold this circumstance to extend significantly further than matters of efficiency and optimization, as something possessing the potential of creating a qualita-

tive transformation, both within architectural design itself and our understanding of it. Through my work – past and present – I see this transformation to stem from an understanding of technology originating in an unruly, material and lived reality. Here

man-made systems are coupled to processes and agencies, which already exist in our surroundings and materials. These become, so to speak, technological – but, on their own terms. I observe, that because we can today describe a surrounding reality through

digital forms of representation, it reveals itself to the design process; as an active dis-

tributed ‘intelligence’ – an agency which reflexively acts back towards the architectural and its processes of design. It could be said that nature – understood in the broadest

possible way – within the design space, can be mediated through descriptions of agencies, which in a reciprocal relation to design describe a seamless technological field.

In the context of architectural practice I believe that this indicates a particular contemporary understanding of what it means to appropriate and apply technology as part

of the process of design. This understanding challenges how technology, in my view,

regrettably and reductively often comes to appear as a means of planning and control. In contrast to such simplification, technology can today assume a role as a reactive and

adaptive means of communication and negotiation between designer, material reality, culture and society. Rather than specifying mechanisms of input and output within

processes, technology can be understood as a means to ask question beginning with an algorithmic and dynamic ‘how’. Consequently, relative to this technological outlook I

observe a generative rather than representational knowledge as the main structuring principle behind the development of a digitally produced architecture.

19


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

It is this generative view of technology, and an architectural understanding of it, that is my overarching interest in what follows.

Towards the Procedural and Generative

Leading up to the commencement of the thesis work – and as my efforts have pro-

gressed towards the above understanding of the technological – the techniques I utilize in order to arrive at design solutions have become increasingly dependent on the

writing of code (i.e. computer programing and scripting). Although a range of digital tools are today available to assist the architectural designer, I observe that within

a critical and experimental practice one always need to go beyond what these tools intend, expect or prescribe. Consequently, I have found it necessary to cultivate an

understanding of the processes by which digital tools and media operate. Associate professor of computer science Michael Mateas refer to such an understanding as procedural literacy:

“By procedural literacy I mean the ability to read and write processes, to engage culturally-embedded practices of human meaning-making and technically-mediated culturally-embedded practices of human meaning-making and technically-mediated processes. “ (Mateas 2005, p.101)

The view attached to this position is that the opposite – procedural illiteracy – leaves

digital practitioners unable to comprehend the essence of computational media, and that it would be naïve to believe that the technical details of digital artefacts are of

no concern in contemplating their significance in a broader cultural context. We need to understand how code operates as an expressive medium. Computer programs, and

their development, are more than components in a project which aims towards “in-

creasing our ability to work with large volumes of information; it also encourages new and different ways of thinking” (Reas McWilliams 2010, p.17).

It has been by developing my own procedural literacy – not as a project of becoming a

computer programmer, but as a means to achieve design freedom – that my perception of a digital design space has become increasingly drawn towards the concept of generative processes. The conceptions above – concerning a technology of distributed intel-

ligence, agency, adaption and reactivity – appear to imply such a perspective. So, while the concept of procedural literacy spans a more general terrain of digital culture I am

20


INTRODUCTION

[1]

1. Drawing made by means of a generative growth process based on simple operations of branching and convergence

21


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

specifically interested in understanding generative processes as meaningful artefacts

in the context of architecture. By themselves and as techniques executed through the formulation of algorithms, generative processes assume the form of technical descrip-

tions, not much else, wherefore an understanding that simultaneously acknowledges and ventures beyond technical execution allows us to investigate a potentially broader significance and relevance of these processes. Generative Design

The concept of generative design should be clarified. While being aware that it is

debatable whether architectural design constitutes an artistic practice, I choose to de-

part from a definition borrowed from the practice of generative art. Here the concept relates to a practice where a system “is set into motion with some degree of autonomy contributing to or resulting in a completed work of art” (Galanter 2003, p.4). Associate professor of information technology Alan Dorin (et al.) suggests that what is shared

by such systems, across media, material, technique and technology are four main constituting components: entities, processes, environmental interaction and sensory out-

comes (Dorin et al. 2012). Through my own experimental work I observe a similar set of components describing an architecturally oriented generative practice, wherefore I believe that this formal model is applicable in the present endeavour as well. Accordingly, entities are constituents that are conceptually “unitary and indivisible, and whose

functional relationships are not typically expressed in terms of internal mechanisms” (ibid, p.244). In general entities are described by properties that are spatial, temporal and

formal. However, the outcome of a generative process might not itself exhibit these properties. Processes refer to “the mechanisms of change that occur within a generative system; they necessarily involve entities that perform operations on, or interact with, each

other” (ibid, p.245). These mechanisms exist independently from the materiality of the system. Environmental interaction refers to an understanding that “all generative systems

operate within a wider environment from which they may draw information or input upon

which to act. The environmental component […] describes flow of information between the generative processes and their operating environment” (ibid, p.246). This can be extended to include the higher order interaction between designer/artist and process. The sen-

sory outcome of a generative process of design refers to artefacts produced from these

processes as selective snapshots, final states or accumulations – as mapping or direct perception of entities.

These four components describe the components of a formal framework for a gen-

22


INTRODUCTION

erative design process, but they do not in themselves address a corresponding and

relevant mode of thinking, which must be assumed necessary in order to direct generative processes towards specific uses, such as design. For a proposal in this endeavour

we may look to a discussion by Manuel Delanda in his essay ‘Deleuze and the Use of the

Genetic Algorithm in Architecture’ (Delanda 2007). This discussion regards how generative design techniques can be used to question deterministic notions of linear causality and replace a traditional design process. But, most importantly the text addresses the

question of how to think in the context of utilising a generative process to address

architectural concerns. What is suggested here is that in order to operate productively and meaningfully, by means of generative procedures, design must depart from think-

ing relative to the concepts of the populational, intensive and topological (ibid, p.408). Populational thinking emphasizes that, “(…) although at any time an evolved form is realized in individual organisms, the population, not the individual, is the matrix for the

production of form” (ibid, p.409). Intensive thinking relates to mechanisms of so-called intensive quantities, e.g. temperature, pressure and speed. These are the mechanisms that produce differences since they drive processes by which the diversity of actual

forms are produced (ibid, p.409). In architecture we might think of these as the coor-

dinated forces which make a building viable. Under the conditions of the previous two

modes of thought, topological thinking is that which substitutes the use of an absolute and metric representation, which loses its meaning in a context like this. Topological

thinking urges design to proceed by means of abstract diagrams describing topologi-

cal invariants, such as the dimensionality of a space, or its connectivity (ibid, p.411). It is through this activity that generative design transcends ‘breeding’. The designer becomes the maker of topological diagrams bearing their signature.

I observe the modes of thought proposed by Delanda to be applicable to the constituting components of a generative design process as suggested above. Thus, entities

can be seen to depart from a populational schema and to be topologically defined; processes are driven by intensive mechanisms, many of which must be identified as

constituents of a productive environment. The definition of generative design assumed by the thesis therefore relates to both a set of formal components and a specific mode

of thinking. However, the above should be read as an introduction or point of origin

for an understanding of generative design, which is continuously developed and extended in what follows.

23


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

Outline

The above outlines my point of departure for the project. In this role it is naturally

simplified and with questions, postulates, hypotheses and context still in need of further elaboration. However, I will briefly summarize. Firstly, I situate the work relative

to a broad framework of a critical exploration of the application of digital technologies in architectural design. Here I observe a contemporary technological potential

of dynamically coupling architectural practice to a complex environment of processes

and parameters surrounding and informing a design. Although such a coupling has obvious practical applications and advantages I am interested in how situations arise

by which planning and control give way to a non-linear and generative procedural

conception of the design space. This leads to a need to understand generative processes

as meaningful artefacts in the context of architecture and consequently how code op-

erates as an expressive medium within a larger cultural terrain; as technologies related to difference.

Contextual Framing

Before addressing the precise inquiry of the thesis in greater detail, I will need to situ-

ate the present work more accurately in relation to the context to which it contributes. In this I observe two overarching tendencies of importance within a contemporary digital architecture to which the work relates: Firstly, a particular interdisciplinary approach to design-practice experimentally exploring the intersection between architecture and digital technology. Secondly, a theoretical tendency characterized by concepts

related to a becoming of form inspired by findings in contemporary natural science

and philosophy; for example as in the case of Delanda above. These two are often seen to coincide in an experimental and digital architectural practice, but I will suggest that

there is no necessary and direct link between them. The theoretical predispositions do not prescribe an experimental digital practice per se and individual instances of this

practice by far exceed what the theories themselves suggest. However, there is never-

theless a significant transfer of ideas from these theoretical standpoints to practice. My point of departure, regarding the contextual and conceptual framing of the thesis, often lies precisely here.

Tendencies of a Critical Digital Practice

I perceive the context of a critical and experimental digital architecture as a range of interrelated but varying focuses followed in practice. These are occasionally pursued in

24


INTRODUCTION

isolation, but more often in combinations; consequently not as clearly distinguishable as I intend to present them here. I do not observe these tendencies as part of a unified

architectural movement. However, they do share a common outset in an attempt to in-

troduce digital technologies, techniques and tools without superimposing them onto a

pre-digital practice. From hereon they individually emphasize more specific interests. I will single out four contemporary tendencies within this field of digital practice to

which I associate the thesis. Below I will summarise the general nature of each for the purpose of broadening the scope of the generative perspective of the thesis, which – at varying degrees – they can be argued to relate to. A further elaboration will be given in

chapter four, where I will return to this topic, and in greater detail unfold the subject. It should be noted that these four tendencies are to be understood as levels within a model for practice, which developed in relation to a much larger field of practice. As

such they represent a model that I construct with the specific framework of the thesis in mind.

Generative Architecture

The generative tendency within contemporary digital practice to which I place the most emphasis is one characterized by borrowing, abstracting and transforming processes from the natural sciences – including computer science – subsequently transforming these into techniques for design. This essentially relies on approaching design

by means of computational and algorithmic techniques. My interest in these methods connects to how they potentially relate design to a radical contemporary shift in the

relationship between technology and knowledge. Technologies have often been ob-

served to follow from scientific theories. But, with a contemporary technologization of science – as well as reality and nature in general – technologies have now become more powerful than our theories; consequently making it possible to do things through

technology, which we cannot fully understand (Cilliers 1998, p.1). As examples here-

of one can consider quantum mechanics, gene-manipulation and pharmaceuticals. Certainly the ability of computer technology to model and simulate the behaviour of complex systems fits into this group as well. It is precisely this ability to operate beyond what human cognition allows which is now seen in certain digital design

practices challenging our notions of control and agency. Here the focus is on process rather than representation, and formation rather than form (Leach 2004). The design

spaces of these practices assumes the form of self-organizing systems often exhibiting emergent behaviour, orchestrated, rather than actually controlled, by a designer. For

25


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

example, this relates design to a concept such as ‘swarm intelligence’; originally referring to the idea that flocks of animals develop an intelligence greater than the sum

of its parts, but in architecture is extended to describe how structure may operate as a self-organising mechanism (Leach 2004).

It is however not my intention to isolate this generative tendency from other trends

within a digital design culture. Designing by generative algorithms can never be done in isolation and they do not constitute a design methodology in themselves. Algorithms are simply rigid rules, followed by a computer, with little resemblance to the

often vague, incomplete, contradictory or interpretable rules made up by a designer. This suggests that algorithmic design approaches may often be required to operate on

only parts of a design and do not constitute a general approach to design(Williams

2004). Because of this limitation I maintain awareness directed towards three addi-

tional tendencies of relevance, which I identify within contemporary digital practices. Procedural Complexity

The first of these tendencies relates to the physical realization of a digital architecture. Not only have the tools by which architecture is designed become digital, in parallel, fabrication technologies for producing components and parts for buildings have

as well. Thus, computation has become inseparable from all stages of the making of architecture (Thomsen 2011), and all stages now potentially interface by means of

being digital. The realization of buildings today can as a result be directed towards

integrating information to form complex parametric assemblages aimed at producing efficiency, overview and complexity-management at different levels of design and fab-

rication. Although this can be utilized with respect to a very conventional architecture, it is also what makes the realization of complex new form-languages and technical experimentation possible. Consequently, increasing design freedom. What is of spe-

cial interest within the present work is the presence of interface between design and making, linking across conventional phases of design realisation (Kolarevic 2008). As I

will return to later in this chapter, the thesis has a special interest in investigating how digital technology challenges the concept of hierarchy in design. Material Agency

Another tendency within a contemporary digital practice develops, more or less, directly from the former. Digital manufacturing and production have not only allowed

26


INTRODUCTION

the realization of advanced geometries and structures, but have also resulted in a renewed interest in the materials by which we build. This links to both how computation allows the simulation of material systems and the precision and control offered by digital fabrication. This interest in materials can be described as an approach to

design operating at the level of the properties and behaviour of material. These are concretely utilized to drive and inform a design process. This occurs either based on

existing material characteristics, as a kind of digital craftsmanship (Thomsen Tamke

2012), or by reproducing properties and behaviour outside of their initial contexts by way of bio-mimicry (Menges 2012). This tendency has developed into positions where

computation in architecture is no longer concerned with producing representation, but as a tool for simulating and mapping performativity. Thus, architecture is no longer

automatically following a spatial language deriving from perspective and projective geometry, but one based on the agency of the material. Cultural Embedment

Lastly, I am interested in practices motivated by the hypothesis that digital technology can act as a catalyst for cultural change, relating to, and affecting external meaning (Rahim 2006). A search for an actual cognitive shift occurring in parallel with

formal, spatial and material qualities arrived at through a digital practice (Goulthorpe 2003). The point of departure is a desire to establish legitimate architectural prin-

ciples aligned with technological change; finding ways to avoid constraining digital

technologies to current expectations of architectural form and to existing practices. The view is that the digital represents a new culture to be inhabited. This culture is already populated by discourses found outside of architecture, but by which we must

engage in order to truly renew the practice of architecture. Important to note is that within these practices we observe the view that technology should never serve as the sole principle of their legitimacy. This necessitates a methodology that can produce

an ‘input’ that lies outside of the technological. A conceptual framing is required by which design acquires a form of resistance against loosing autonomy to an instrumental reason, wherefore concepts from outside such reason are required (e.g. the practice of R&Sie[n])(Corbellini 2009).

In summary the context I observe in contemporary digital practice, and to which I re-

late the present work, challenges the role of the designer, the nature of representation, how we build and how we affect and become affected by a contemporary culture. What

27


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

I will propose is, that a significant shared characteristic exist across these tendencies

in the form of a relinquishment of absolute control and the presence of non-linear or

cyclical dependencies across, and in relation to, the design space. Although this does not suffice to describe these trajectories of contemporary architectural practices as generative design, it is my hypothesis that they can all nevertheless be conceptualised and approached relative to a generative perspective on design.

Constructing the Inquiry

A generative approach to design – and the procedural as well as technical understand-

ing associated with this type of practice – implies a certain mode of engaging reality. By coupling processes of design to agencies displaying autonomy, by acting relative to

systems exhibiting a complexity beyond understanding and by extending our technical repertoire though digital computation we need to consider our perception of the designer. In this context, if one holds a view of the designer as an authority controlling

and planning a process of design, then this view must be challenged. This as well per-

tains to a mode of considering technology as an applied science, where a certain input produces an expected output (e.g. Lyotard 1984). This view cannot persist in a genera-

tive model for design, because prediction is simply not meaningful due to the locally determined and complex dynamics of such systems. Even outside the perspective of

the generative it can be argued that methodology cannot be something that prescribes

to a design process a stringent procedure, which systematically collects information,

establishes objectives, and computes the design solution (Gedenryd 1998). However, a generative model of design would seem to further emphasise the view that design is not planning.

In the following I would like to put forward two hypotheses. These address how a

design practice operating by means of algorithmic and computational techniques –

in the form of a generative design space – might contribute to a contemporary and digitally produced architecture. The first hypothesis relates to a notion of the overall structure of a design process, while the second relates to the nature of control within it. Thus, I will suggest that through the furthering of a generative approach to design: 1. A hierarchical structuring of the design space may be exchanged by a network

of entities exhibiting degrees of autonomous agency; consequently suggesting structure as self-organising

28


INTRODUCTION

2. Planning may be exchanged by the mechanism of reflexivity In this I do not necessarily posit that hierarchy is the prevailing structure by which a

design space is ordered today. What I do suggest is that hierarchy is a convenient and

effective way to deal with the complexity arising within any process of design. Thus, the utilization of a hierarchical structure presents an immediate and obvious solution

to the complexity that I observe in both my own work and that of others – if noth-

ing else, as a momentary ideal. Historically this has, in certain instances, appeared as a methodological necessity. Christopher Alexander observes in ‘Notes on the Synthesis of Form’ (Alexander 1964), that the designer constructs hierarchies to deal with the

highly amorphous and diffuse conditions of the problems they are confronted with. Here the form-making process is pictured as “the action of a series of subsystems, all in-

terlinked, yet sufficiently free of one another to adjust independently in a feasible amount of

time” (Alexander 1964, p.43). Although I will nuance this view below, in the present context a notion of hierarchy would appear to be contested by the idea of a generative approach to design, since we here relinquish control, as well as an understanding of the structure of a design process as something pre-existing.

Returning to the above hypotheses these are in fact highly interrelated since the notion of planning implies hierarchy, and reflexivity may be perceived as a property of

autonomous agency. A correlation between a model of structure and a model of control seem to exist in both point of departure and the proposed target. As such I observe the second hypothesis with the most importance. Thus, by the end of this section it is

my intention to accentuate the significance of the mechanism of reflexivity as defining the overall problem domain of the thesis. As a stepping stone towards this end, I will

however briefly begin by addressing the key concepts of the first hypotheses; hierarchy and self-organisation. As observed here these concepts appear as simple oppositional

constructions. Though, within their oppositional schema I nevertheless see something

which point beyond it; towards something more ambiguous and complex than what its simple construction might at first convey. Hierarchy

Evolutionary biologist Richard Dawkins defines a hierarchy as a set which satisfies

two fundamental requirements. Firstly, that there is no element in the set which is superior to itself, and secondly, that there is one element in the set, which is superior

to all other elements in the set (Dawkins 1976, pp.9-10). The first requirement pre-

29


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

vents circular relationships; the second ensures that every hierarchy departs from a

single point. Conversely Herbert Simon proposes that it is only in a general use of the word that it designates systems in which each of the subsystems is subordinated by an authority relation to the system it belongs to (Simon 1996). Simon refers to these as

formal hierarchies, whereas his own definition refers to all complex systems analysable into successive sets of subsystems. In the context of the thesis hierarchy will however refer to the more narrow definition given by Dawkins, while other terms will denote complex systems of a different nature.

According to Dawkins there is good reason to operate within hierarchical structures (Dawkins 1976, pp.16-18): Firstly, hierarchical classification is a vital convenience

for everyday life since this is a universal means of organizing information for easy ac-

cess. Information retrieval requires systems under which we can subordinate things; addresses, zip-codes, alphabetical ordering etc. Secondly, dividing assemblies up into

sub-assemblies are more efficient than assembling from an unsorted set of components. A hierarchical structure allows a division of a system, so that local and global

decisions can be made appropriately to their position within the system. Lastly, ordering components in hierarchies reduces redundancy because elements can be arranged so that specific functionality can be assigned only where needed. In a non-hierarchical

structure a component’s role, within a system, would require it to be highly general, and therefore in danger of being unnecessarily. Self-Organisation

While hierarchy can be understood as a useful tool in an endeavour to address complexity, the principles of self-organisation represent a contrasting model for this same task.

The dynamics of self-organisation are a general property of complex systems (Gros 2008). When we talk about the structure of a complex system we refer to the internal

mechanism developed by the system to receive, encode, transform and store information, as well as to react to such information. Given certain conditions self-organisation

in a complex system allows structure to evolve without the intervention of an external

designer or the presence of some centralized form of internal control (Cilliers 1998, pp.89-90). It develops a distributed internal structure through the processes it gives birth to. Structure is here neither a passive reflexion of the outside nor the result of active pre-programmed internal factors. Instead structure is the result of a complex

interaction between the environment, a present state within the system and the history

30


INTRODUCTION

of the system.

Self-organising structures can be viewed as decentralized problem-solving systems, comprised of many relatively simple interacting entities (Bonabeau et al., 1999). This relies on the idea that a group of agents may be able to perform tasks without explicit

representations of neither environment nor other agents, and where planning may be replaced by reactivity (Carranza Coates, 2000).

The advantage of using self-organisation as a tool to solve problems is that it presents a flexible and robust way of doing so. It is flexible in that it allows adaption to changing

environments and robust through the ability to function even though some entities may fail to perform their task. The disadvantage of using self-organisation to solve problems resides in the bottom up approach to creating such systems. Here the paths

to problem solving can never be predefined but are always emergent and result from interactions among entities themselves as well as between entities and their environ-

ment. Therefore, using self-organisation to solve a problem requires precise knowledge of both the individual behaviours and what interactions are needed to produce a desired global effect (Bonabeau et al., 1999).

The significant characteristic of the concept of self-organisation is that it is a principle for navigating within complexity without hierarchy. According to Sanford Kwinter

our reality consists of systems that are so extensive and complex that they cannot be

described through an absolute representation (Kwinter 2001). Should we accept his

claim, then articulating a design problem as a self-organizing system might be a candidate solution for dealing with this challenge?

Problem Domain: The Mechanism of Reflexivity Loss of Hierarchy

Dawkins definition of hierarchy resembles what Deleuze and Guattari call arborescent systems (Deleuze Guattari 1987). They define arborescent systems as structures

where elements only receive information from a higher unit, and only receive a subjective affection along pre-established paths. Deleuze and Guattari raise a critique against

such systems due to the fact that they are incapable of reaching an understanding of the concept of multiplicity. This concept can, at the present time, be defined as the

virtual structure underlying a process of becoming, without which novelty and creativ-

ity would be impossible (see chapter 7). By structuring our thinking hierarchically we limit the possibility of change and difference. Instead of the hierarchical/arborescent

31


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

Deleuze and Guattari propose the use of the alternative system of the Rhizome to think in relational terms. We may understand the structure of such a system by envisioning a map that is continuously produced and constructed, that is always detachable, connectable, reversible and modifiable, and has multiple entryways and exits.

In light of the fact that there are obvious advantages in hierarchical structuring it is worth noting that Deleuze and Guattari acknowledges that there are “knots of arbo-

rescence in rhizomes, and rhizomatic offshoots in roots” (Deleuze Guattari 1987, p.20). Equally Dawkins sees no reason why the position of hierarch (an authority under which we subordinate other elements in a hierarchy) cannot shift, for example within a networked hierarchy (Dawkins 1976).

For me this suggests a possible positive relation between something hierarchical and its apparent opposite, by which we might not have to subscribe to an oppositional

schema but can operate relative to a productive conflict. Consequently a potential loss of hierarchy within an architectural design process, suggested by a generative perspective, should not be thought as an absolute static state, but rather as a condition or possibility of becoming lost. Regardless of how complex and self-organising a system for

design might become, one must assume that even in the process of generative design

things do momentarily solidify in the form of something imposing its authority on the design process. We do design with a task or intention in mind. As defined above a

generative design system must be set I motion by a designer, placed within an appro-

priate environment and its entities described. By means of diagrammatic instructions certain goals are imposed upon the system and its end result is required to represent

a certain degree of unity. And yet, at the same time the agency of a generative system and its constituents will at times assume their own autonomy. Occasionally it will

propagate and unfold with a degree of independence to which the designer becomes

a mere spectator. These processes will unavoidably yield unexpected results to which

corrections to their constituting elements must be made; in this way relating technol-

ogy to difference. Thus, the apparent conflict between the pragmatics of hierarchy

and the dynamics of self-organisation might in fact be a productive conflict driven by the dynamics of differences. I will suggest that this observation can be contained by

the concept of reflexivity. In the words of literary critic N. Katherine Hayles reflexiv-

ity “[…] is the movement whereby that which has been used to generate a system is made,

through a changed perspective, to become part of the system it generates” (Hayles 1999, p.8). Reflexivity imposes a shift away from a representational approach to reality towards

a performative. Rather than thinking of the world as consisting of facts and observa-

32


INTRODUCTION

tions, one can proceed by the idea that it is in fact filled with agency (Pickering 1995). Consequently, in the context of the thesis I understand reflexivity to designate a productive conflict between, on the one hand, control and planning, and on the other, the

autonomous agencies of generative processes. Reflexivity is the mechanism which set the design space in motion relative to fluctuations within a fleeting condition of a loss of hierarchy. Aim

The aim of the thesis is to contemplate a problem domain attached to the significant mechanism of reflexivity, developed relative to a generative approach to design. Here my overall goal is to produce a map of the terrain defined by this concept – in this

context – by which a practice might navigate. I observe this as the core issue of the thesis. Rather than making the present work a project which attempts to present tech-

nical solutions or definitive answers the ambition is to establish a productive surface

from which a variety of solutions might arise and different questions might be probed. The project does not seek to arrive at a final answer to an overarching question, nor

a general model for design. Rather, the ambition is to stage the conflict between, on

the one hand, the pragmatics of hierarchies and planning, and on the other, the unruly

dynamism of self-organisation and autonomous agency; this, by contemplating the difference-producing concept of reflexivity – as well as questions existing in parallel

to it – from the perspective of a critical digital design practice. Thus, if a question does exist, it is to ask:

How can one approach the construction of a framework for a digital design

practice, in which the hierarchies of planning and control are productively con-

fronted by the orchestration of self-organising agencies, within a technological field structured by differences produced through a mechanism of reflexivity?

By defining a core issue, identifying a problem domain and posing a question in this

way, I aim to signify that with the present work I do not attempt to mimic a scien-

tific practice. There is a difference between what I do and what a scientist does, even though I do relate the work to a certain technical outset. This affects how knowledge is produced within the thesis. Rather than attempting to produce new knowledge by answering a fixed question the thesis seeks to map a terrain defined by the observed

core issue of reflexivity. This map – which I will refer to as a framework for design – is

33


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

methodologically established in exchanges between discourse and practices of making. Thus, the problem domain defines, and is investigated by, movements through both a conceptual and practice oriented field, where parallel paths are defined by parameter

that do not attempt to answer a shared pivotal question. The intention is to offer pro-

posals rather than answers and to identify possibilities rather than prescribing courses action. I will elaborate further on methodology in the following chapter.

Contribution: A Framework for Design

As a means to address the above intention and as a consequence of its specific for-

mulation my goal is to conceptualize and develop a framework for a digital design practice. It is this endeavour that characterises the practice of the thesis and it is the

form I apply to the result of the thesis. In the context of the thesis; by framework I refer

to a productive field supporting and enclosing a set of interrelated elements, which

can be used as a means for assigning a certain directionality to a practice of design. The framework is perceived as a kind of map designating both points of orientation and areas of movement. The aim of this specific framework is to contribute to current

developments of architectural practice by which computational and algorithmic tech-

nologies are utilized to arrive at new procedural and generative perspectives for design. Here the framework serves the purpose of offering an overall form for the thesis, by which I can investigate the focus circling the concept of reflexivity, which directs the

work. In this matter my overarching interest is to offer suggestions as to how the

technologies and concepts in question might contribute a difference to a digital and experimental practice of architecture and to assign a distinct characterization for the nature of this difference. This links notably to questions concerning the use of code as part of the technical repertoire of the designer.

The framework also offers a means to align the present work to a range of models, theories and concepts which the research shares with other practices. This includes

models of societal/cultural structures, aspects of current digital design trajectories, cybernetic theory, a critique of a simplified absolute representation and theories of

becoming. In this way the framework acts as a means to interface the practice of the research with an outside larger than itself. Limits

The framework seeks to specifically offer directionality to what I refer to as a critical design practice, as a means of delimiting the scope of the project. In this regard the

34


INTRODUCTION

word critical and the concept of design should be addressed for clarity

Herbert Simon offers the definition of design-activity as the devising of “courses of

action aimed at changing existing situations into preferred ones” (Simon 1996, p.111). A more recent proposal made by associate professor of design science Paul Ralph

and professor of information technology Yair Wand offers a definition of design as: “(noun) a specification of an object, manifested by an agent, intended to accomplish goals, in a particular environment, using a set of primitive components, satisfying a set of requirements, subject to constraints; (verb, transitive) to create a design, in an environment”

(Ralph Wand 2009, p.107). Overall I tend to agree with both descriptions. Simon’s designates design as something aimed towards a qualitative difference, while Ralph

and Wand’s more formal definition appears conceptually compatible with the generative perspective of the thesis; this, by emphasising the concepts of agency and environ-

ment. However, in light of situating the present work in the context of architectural practice I will in addition assume that design always include the development of form and visual representations, which is not explicitly stated in the two definitions. Thus

within this project I understand design as defined by Ralph and Wand with the ad-

dition of a supplementary requirement of producing a qualitative difference aimed at the creation of form.

Regarding the concept of a critical design practice, what I refer to is a specific ‘labora-

tory’ situation; that is, an exploratory and experimental practice. As to the nature and target of the ‘critical’, this points towards the experiment and the experiment’s ability

to create a difference. When I write about design practice throughout these pages it is this kind of design practice I reference. However, I do keep other types of practice in

mind when considering the relevance and practical implementation of the results of

the work. In this way the framework is understood more as a supplement than substitute to a more general practice of design. Discovery as Method

With an emphasis on the concepts of difference, the experimental and exploratory, I do not intended to describe the framework as a pre-existence, somehow necessitated

and stabilized by a context defined by others. Rather, my ambition has been to discover this framework, as an emergent structure, and to suggest possible paths for it

to follow. This has triggered a certain methodological standpoint. For me, discovery implies situating oneself consciously within a terrain of knowledge. Observing from within – learning, predicting, mapping, connecting and – inevitably – inventing as

35


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

well. Thus, the work could be described as neither objective nor subjective, but relative. That is, acknowledging the presence and context of the observer, subordinated by

neither institutions nor idiosyncrasies. I borrow this kind of positioning from the field of cybernetics, where the agency of the practitioner must always be situated within the

area of inquiry (Heylighen Joslyn 2001). I will discuss and relate this theoretical field to the framework in chapter five.

Overall Construction of the Framework

Within the terrain explored by the thesis, the construction of the framework has been approached in a dual movement between discourse and making. In this way the dif-

ferent parts of the thesis find their point of departure, sometimes in procedures of making, and sometimes through the contemplation of ideas, concepts or the positions

of others. From this outset practices of writing and practices of making inform and interact with each other forming productive components related to the unfolding of a

digital design practice. As these parts develop they reflect back on the terrain defined by the present problem domain – as a kind of mapping, sectioning or re-configuration – and thus the framework appears. Not as a description of the context but as a

contribution often exhibiting productive conflicts within the material. As such, it is important to state that the framework should not be interpreted as being a proposed

methodology – as something prescriptive. I use the concept of framework precisely to

designate a space, or structure, in which a considerable degree of freedom and individuality persist, so as to allow continuity, multiplicity and further development. Constituents and Target of the Framework

What does the framework consist of ? It is an assemblage of guiding components from which design spaces can be constructed, by suggesting directions, delimitations and

relations. Thus, it contains elements pertaining to methodology, technology and ideol-

ogy without being either. Its purpose is not to unify, but to set in motion. The target of

this motion relates to the project’s involvement in the use of digital technology. What is a significant focus of the project is to observe how the consolidation of a techno-

logical space, on the one hand, and a design space on the other, facilitates a potential qualitative transformation of a digital architectural practice. That is, to speculate on how the technologies of interest might effectuate a genuine change within design –

becoming something other than a prosthesis. When I intend the framework to direct design towards a target, it is to this very end.

36


INTRODUCTION

Questions

Although my aim is this open-ended structure of a framework for a design practice, the terrain which I explore does suggest and require the probing of questions. Moreover these serve to designate components within the framework and resonate with

the structure of the thesis itself as offered below. The most significant amongst these questions are:

What is the model by which to address the application of technology relative to a generative

approach to design? Since the mechanism of reflexivity implies a non-representational

circular approach to reality, technology cannot follow the model of an instrumental reason, which must be assumed to require a representational form of knowledge and a linear relationship between tool/technique and target. What other model may guide the application of technology?

How can a generative approach to design be positioned within a broader context of a digi-

tally produced architecture? In order to represent a relevant contribution to architectural design, a generative approach to design must be aligned to a relevant context. This

also offers a tool in an endeavour to understand generative processes as meaningful components within a larger field of architectural practices.

How can a practice of writing computational procedures be understood as an approach to design? In order to fully implement a practice of digital coding as part of the repertoire

of the designer we need to consider this activity in relation to questions of designmethods and practice.

How are we to understand the concept of representation within a generative perspective on

design? A conflict appears to exist between, on the one hand, a generative, reflexive and non-representational approach to design, and on the other, architectural design as a practice engaged in the production of representations.

What is the nature of a generative design space as a whole, and how might this entity be described and become operational? Combined with the question of representation the el-

ements constituting a generative design approach challenges how the workings of the

design space are orchestrated. It requires a structure by which to situate the designer, (autonomous) entities, (generative) processes and the construct of an environment.

37


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

Thesis Structure

I perceive each chapter of the thesis as an element designating a certain area of im-

portance within the suggested framework for design; consequently, assuming a cer-

tain degree of autonomy by which they develop. This resonates with the image of the framework as a kind of map. I use this image to indicate that the set of elements addressed and proposed – in the form of the individual chapters – do not assume the

appearance of a list indicating a progressive subordination, but are to be considered as

levels within an open-ended model. In this way I have been cautious of establishing too strict a linear causality across the chapters. This allows me to re-address the findings of the thesis within the concluding chapter as a process of surveying the proposed territory following different trajectories.

The thesis is divided into three main sections. The first section (chapters one and two) presents the preliminaries required to understand the overall framing, context and approach of the work. The second section (chapters three to five) comprises a

movement directed towards the external relations by witch the framework is aligned

to a broader context of ideology, theory and practice. This movement is intended to describe a transition from a general discussion of technology towards the specificities of the thesis. The third section (chapters six and seven, as well as experiments one, two

and three) discusses the internal workings of the framework through reflections on elements of a digital design space and the findings of the major experiments conducted

throughout the execution of the thesis. The texts regarding the individual experiments assume a slightly different form than that of their surrounding chapters. Firstly, they

are contemplated from their own individual perspectives, and thus unfold from more locally restricted research problems than the chapters. Secondly, they include descrip-

tions of specific techniques, procedures and pragmatics, which are elements generally not discussed in the chapters. I include these to allow the reader an understanding of

the concrete form of work that the practice of making within the thesis assumes. The thesis is concluded with a summation and critical reflection of its results.

Structure: Preliminaries

Chapter 1 - Introduction

The aim of this chapter is to present the overall contextual framing of the research, its point of origin, key interests, hypotheses and core issues. The main purpose is to

38


INTRODUCTION

formulate how the project intends to describe and construct a framework by which

a generative design practice – operating by means of algorithmic and computational technologies – seeks to implement an experimentally focused practice operating relative to the notion of reflexivity. Chapter 2 – Methodology

The aim of this chapter is to describe the method by which the thesis has been developed. The key discussion relates to how the construction of the framework is devel-

oped relative to a dual movement of discourse and making. Additionally the chapter defines how the thesis understands architectural research, appropriates and borrows from the positions of others, perceives the role of practice and the nature of the experiments of the thesis. Part 1

Chapter 3 - Technology beyond Optimization

The aim of this chapter is to present an ideological and conceptual outline for the ap-

plication of technology in the context of an experimental and digital design practice. This relates to the research hypotheses of control and structure, seen in a cultural con-

text. The purpose of the chapter is to develop an understanding of the broad cultural/ societal context to which the thesis relates; this with implications for the project as

both framework for design and research with a focus on questions of technology. As

the title suggests the position developed in this chapter argues for an understanding of technology which does not subscribe to that of an instrumental reason alone. Chapter 4 – A Manifold Digital

The aim of this chapter is to consider the specific context of a digitally conceived

architecture to which the research and the framework relates. The purpose of this investigation is to position the work relative to this context, indicate relevant discourses

within it and to point towards where the work itself, reflexively, contributes to the field. Thus, implicitly describing the outer contours of the framework for design through a discussion of the positions to which it is aligned. Chapter 5 - A Cybernetic Approach

The aim of this chapter is two-fold. Firstly, the chapter shows that the field of cybernetics can be utilized to define the role of the designer. It supplies a conceptual model

39


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

by which we can choose to understand a system for design. Although this relates to

an understanding of design in general, key aspects of cybernetics are particularly ap-

plicable for conceptualizing across the experiences of a generative practice of design. Thus, offering the outline of a conceptual framing of the specific design spaces relevant to the thesis. Secondly, the chapter shows that cybernetics presents a number of useful

models in a conceptualisation of the notions of agency, representation, environment and the important mechanism of reflexivity. Part 2

Experiment 1 – Lamella Flock

This experiment explores a situation in which an agent-based solution space is used to establish a design practice characterized by the simulated agency of material components, under the ‘steersmanship’ of the designer. Chapter 6 - Representation and Agency

The aim of this chapter is to address the topic of architectural representation in light

of the framing suggested by the previous chapters. My purpose in this chapter is to conceptualize agency as a mechanism introduced into representation – a mechanism

that allows representation to assume a reflexive relation to that which is represented. In this way I attempt to refrain from subscribing to an oppositional schema, by thinking representation as a potential vehicle of agency. I propose to envision that which

constitutes representation within the framework of the thesis, as a device for steering the space between design and making in the form of a map comprised of concrete form and abstract notation.

Experiment 2 – Porous Ascend

The experiment investigates how algorithmic and generative techniques allows for the

utilization of complex, and by other means inaccessible, ways of devising the schema by which we arrange the parts of an architectural object. Chapter 7 - Environment and Interface

The aim of this chapter is to address and expand upon the design space suggested in

chapter 6. I will make two proposals in this endeavour. Firstly, that the underlying field of a design space becomes conceptualized relative to the concept of environment and that this concept in fact dissolves the divide between context and object – merging

40


INTRODUCTION

environment and design space. Secondly, that this notion of a design space becomes operational through the concept of interface. Experiment 3 – Scatterings

This experiment explores the idea of the design space as an environment in which the designer operates at the edge of control and description – with and in time. What

evolves within this environment does so without an underlying diagram structuring

the configuration of elements. This ability lies solely within the animated components of the experiment themselves, which assume an investigative type of agency. Ending

Chapter 8 - Framework

The aim of this chapter is to close the thesis – not in the form of a point of termination, but as a situation for new departures. The chapter follows two trajectories. Firstly

it summarises the thesis by means of accounting for momentary generalisations and propositions applicable to a wider context of digital practice. Secondly, is discusses a territory of further investigation through a discussion of how a practice might appear as defined by the total map of the framework. Videos

Supplied with the thesis is included a disc of animated sequences collected throughout

the project execution. These videos accompany the texts regarding the key experiments of the thesis, as well as a number of minor experimental probes. These latter appear as

parallel illustrations throughout the thesis. The animations are named with reference to their corresponding chapters.

41


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

Preliminaries

Part 1

Introduction Methodology

Discourse

Technology beyond Optimization

An Manifold Digital A Cybernetic Approach

Making

3D cellular automata probe

Generative Drawing

Non-Linear Systems Dynamics Probe

Probes

[2]

2. Thesis structure. The texts of the thesis organized relative to their immediate relation to the modes of discourse and making by which the work has progressed, as well as indications of minor experimental probes present within the texts. For further elaboration please see the following chapter.

42


INTRODUCTION

Part 2

Ending

Representation and Agency

Environment and Interface Framework Experiment: Scatterings

Fluid simulation Probe

43

3D Self-Assemblage Probe

Experiment: Porous Ascend

Programmed drawing: 3D Flow Field Probe

Generative Reciprocal Beam Patterns Probe

Experiment: Lamella Flock


2 | METHODOLOGY


METHODOLOGY

The challenge has been to ultimately arrive at a position through which a framework

for a specific design practice, departing from computational and algorithmic technologies, could be suggested. The aim of this chapter is to explain the components of my

methodology and to show how they are utilized to form and propose an understand-

ing of such a framework. The overall purpose of the methodology is therefore to offer a structure which allows the diverse sequences of events, activities and observations

– representing the work of the thesis – to be directed towards accomplishing this goal. The question of methodology is essential because it brings attention to the relationship between technology, practice and knowledge within the work.

By means of introduction, the approach to the work described in this chapter can be envisioned as an investigative, experimental and inventive practice, unfolding relative to three interrelated methodological elements: an overall operational structure

between making and writing, the framework for design itself and a series of guiding components describing productive points of orientation by which this framework might steer.

The operational structure is considered to be the point of departure for the method-

ology. It can be described as two interdependent modes followed through the work, operating respectively to produce discursive and non-discursive elements. One mode

is directed by an argumentative pursuit. It has as its main objective the elaboration and relating of knowledge and concepts across the complete horizontality of the thesis work. Another mode is directed by an architectural design-pursuit and has as its main

objective the preparation of a space for new design procedures. It is in a dialog be-

tween these two modes of operation that the development of the proposed framework for design is pursued – and relative to which it should be understood.

This framework was discussed in the previous chapter as an element representing the

form and result of the thesis; a heterogeneous collection of elements reminiscent of a map – a density which is build up relative to the thesis. The framework for a design

practice should, importantly, be understood as more than a passive element represent-

ing the goal of the work; something positioned as a last element in a linear progression of activities. Rather, it has developed through a necessary conflict between the discourse and making of the methodology, where the framework can be perceived as an

element which evolves through a continuous negotiation between itself and an emerg-

ing conceptual, technical and productive context. This relates to the last element listed

45


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

above, described as a set of components representing productive points of orientation by which the framework might steer. These elements are similarly constructed with the

overarching dual movement in mind, as tools directed at designating the outer contours, interfacing surfaces and productive forces of the framework, intersecting both

discourse and making. Thus, the construction of these components is both part of the structure of the framework itself and extends beyond it, so as to supply perspective and

coherence. They produce underlying models that define what the framework is and

isn’t, and does and doesn’t. As such they tend towards the normative while refraining from becoming polemic or moralizing.

In the following I will in greater detail explain and validate this methodological construction. Firstly I will account for the operational structure that sets the research in motion between discourse and making. Secondly, I will turn towards the framework for the specific design practice that is the topic of the research. Since I have already

addressed this subject in the previous chapter, I will here focus on elaborating the

framework as a methodological element. Thirdly, I will address the structure of the guiding components; that is, the singular mechanisms which animates the methodol-

ogy as well as the proposed framework. Lastly, and as a consequence of the discussions that will have preceded it, I will elaborate on the nature and definition of the concept of practice in light of the proposed methodological structure.

Structure of the Methodology Operational Framing

The overall point of departure for the methodology – as two parallel pursuits centred respectively on discourse and making – were chosen to accommodate three problems observed within architectural research; firstly a question of the relationship between

architectural research and science, secondly a question of determining a standard by which to evaluate a research outcome, and lastly a question of the relationship between theory and practice in architectural endeavours.

Contemporary design research has evolved beyond a purely technical interpretation

of design, concerned with techniques for achieving goals, towards including an exploration of its own epistemology (Cross 2007). In this regard architectural (design)

46


METHODOLOGY

research constitutes its own set of epistemological procedures. Whereas research in the natural sciences is traditionally modelled on hypothetical-deductive reasoning and the

probing of empirical evidence for processes of falsification/verification (Giere 1997), architecture, and architectural thinking in particular, partly develop from a rhetorical and argumentative history ( Johnson 1994). Furthermore, architecture includes creative processes that contribute to its making. These make distinctions between objective theory, in a traditional sense, and positions of personal interest (political, aestheti-

cal, ethical etc.) problematic. Instead of proof in a conventional sense the research has therefore endeavoured to posit another method of evaluation that takes into account

the diverse areas by which architecture is seen to develop. Here evaluation is based on observing how results and findings qualitatively contribute, change or expand existing

situations within an architectural practice. In this matter I delimit the work to specifi-

cally address architectural practice in an experimental situation. It is from within this

situation that it is contemplated and from which it should not be separated unless otherwise stated. That is, I do not claim to make direct contributions to a generally applicable model of architectural design practice.

In order to understand this position I will make use of an observation made by Deleuze and Guattari, who identify two formally different conceptions of science that

ontologically form a single field of interaction; royal and nomad science (Deleuze Guattari 1987, pp.367-369, 373-374). Royal science represents top-down, institution-

alized and conventional science, while nomad, or minor science, is a locally determined scientific field linked to both technology and practice. Royal science continually ap-

propriates the contents of nomad science, while nomad science continually cuts the contents of royal science loose in the vicinity of a constantly shifting borderline. Royal

science is often the one that assumes primacy, “when this primacy is taken for granted

nomad science becomes portrayed as a prescientific or parascientific or subscientific agency� (Deleuze Guattari 1987, p.367). I observe that when architectural research takes the primacy of royal science for granted, or defines itself as such, it positions itself relative

to a situation where it becomes impossible to understand the relations between science, technology and practice. These relations can only be explored by nomad science

because it is a field of enquiry that refrains from distinguishing between intellectual and manual work. Conversely royal science cannot explore these relations precisely

because it divides work up into theoretical and practical. In this, royal science imposes form on matter, while nomad science draws form from an unprepared heterogeneous

47


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

matter laden with singularities, opposing a form/matter duality. Also, nomad sciences

do not allow the scientific to assume autonomous power or development because they subordinate all their procedures to conditions of intuition and construction, while

royal science isolate all operations from these conditions, making them intrinsic concepts, or ‘categories’. The nomad sciences quickly overstep the possibility of calculation, i.e. determination, because they reside within a complexity that exceeds the space

of reproduction. Solutions to occurring problems arrive from activities that rely on

being dependant on the surroundings in which they take place, i.e. they are situated and tactical.

In an effort to stabilise findings, and to exhibit communicable and relevant knowledge, the thesis necessarily assumes a degree of consolidation between these two sciences. However, the present research does tend towards the nomad sciences rather than the royal and this has implications for how it unfolds. First, evaluation depends more on

identifying the complex environment to which the research might contribute and less

on identifying institutionalized general standards by which to measure the outcome. Second, if we follow Deleuze and Guattari’s argument the research cannot operate

relative to a division between theoretical and practical work, since the research is interested in a shared field between science, technology and practice. Instead the con-

cept of practice assumed within the project is manifold rather than singular. The key understanding in this regard is inspired by the position of Stan Allen, where theory

and practice ceases to be oppositional modes of architectural reflection (Allen 2000). Allen observes that theory in architecture has traditionally served as a unifying func-

tion, representing an ideological framework supporting an otherwise fragmented field of disparate demands. Practice has in this appealed to overarching grand narratives

existing outside of practice and building site, expressed in mediums foreign to these. Conversely, in situations where practice consciously renounces theory it realizes unstated assumptions of a theoretical nature anyway. In the end, both approaches reduce design to the implementation of rules set down elsewhere. Instead we should consider

theory and practice as something produced within a definable space, by active conscious subjects. That is, again as locally determined. Here nothing called practice or theory exists, rather, the development of architecture fluctuates between practices of

writing, and material practices. The former being critical, discursive and interpretive, the latter engaged in the production of new objects and new organizations of matter.

48


METHODOLOGY

Productive Conflict

This last idea of fluctuating practices is a convenient image for explaining the conception of the overall methodological framing of the research, but also for positioning the

research results and their corresponding problem domain. Although the work of the thesis in the present form is documented through text (with illustrations) the work

has developed precisely in a dialog between practices of writing and practices of making. It is as a product of this dialog that the research results should be observed, and

it is within this dialog that the core issue of reflexivity is being formulated – neither

as shedding light on a conventionally posed academic or scientific problem, nor partaking in the development of a subjectively defined design practice or poetic. This also

indicates that the framework for a design practice, pursued by the present research en-

quiry, materializes within a productive conflict between writing/discourse and making. Writing has served the purpose of archiving, appropriating and evolving knowledge

and concepts from other architectural (writing) practices, as well as from reflections conducted outside of the realm of architecture. Making has served the purpose of synthesizing and testing techniques; appropriating technologies; transforming ideas from

the practice of writing into procedures for design; and certainly also offered questions to be addressed through the discourse of the writing. Here neither practice is seen to

fully describe the framework by itself, nor do they constitute a symmetrical relation, where elements within one always correspond, prescribe or respond to elements in the other. In fact, the content of the writing, and that of the making, occasionally exhibit

degrees of discrepancy. I observe this situation as productive rather than problematic, in the sense that the awareness of conflict can be used to identify new investigations

and iterations within the research. Instead of attempting to offer absolute answers, the research seeks to articulate new points of departure. This position is present both as

part of the process leading up to the composing of this document and observable in the end result, where certain conflicts have been allowed to persist.

The Framework

The fluctuation between discourse and making represents the most general layer of the methodology. It serves to define an overall structure from which the work has been

developed. As stated above, within this operational structure a more specific model is constructed, namely the framework for a design practice pursued by the research. In

the introductory chapter I offered a description of this element as a productive open-

49


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

ended field – reminiscent of a map – supporting, describing and enclosing a set of interrelated points of orientation and areas of movement. These are thought as a means

for productively steering a practice of design. Each chapter within the thesis serves the purpose of contributing findings from the research to the establishment, description

and development of this framework. In this way it serves a double purpose; as a tool for referencing the findings of the research and as the result of the research itself. As

such the framework can be perceived as the outcome of the research enquiry in the form of suggestions as to how a practice of design might unfold, rather than as a proof of a (‘scientific’) proposition.

To explain how I define and understand the nature of the framework as a method-

ological element I will briefly return to the writings of Deleuze. Here a discussion of

the methodology of Foucault – his archaeology – serves as a useful insight (Deleuze 2004). In Deleuze’s words Foucault establishes strata in which a double rhythm takes

place; between things and words, content and expression, the discursive and non-discursive. This is of course reminiscent of the assumed movement between discursive processes

and processes of making above. Any historical formation sees and renders visible what

it possibly can, delimited by its own conditions for visibility, and as well articulates as much as possible, defined by its conditions for making statements. Here knowledge is

produced through practices constituted relative to many different thresholds (scien-

tific, aesthetic, political, etc.), where the movements and positions of these thresholds

are what distinguishes one strata from another. The challenge of understanding strata is that even when a statement exists it might often be concealed within the structure in which it exists. Something similar applies to the visible, where an object cannot be

understood without also knowing the conditions through which it is made possible. Architecture is for Foucault not only an arrangement of material and a combination of qualities, but a way of displaying and concealing (Deleuze 2004, p.75).

Although the thesis does not attempt to map, or analyse, a historical or contemporary

terrain this image of an archive of elements, some in the form of statements and some

in the form of visible objects, serves the purpose of explaining the role of the thesis’ key element – the framework for a design practice. The components, produced by the research, define a field of interrelated parts – and the spaces between them – which in concert allow an understanding and probing of the problem domain defined by the

thesis. They are relatable in light of the fact that they are made to belong to the same referential map. But, within the thesis it is important to note that they also create this

50


METHODOLOGY

map. That is, the map does not pre-exist the thesis, but is developed by inventing and

discovering the discursive and non-discursive elements that populate it. What we ar-

rive at through this process is a model (the framework) which creates the conditions

for a practice and as such is not a representational but an operational model. It allows for a practice.

Directional Structure: Points of Orientation

I will now address the individual elements which manifest relative to the space of the framework as a result of solidifications within the movements between discourse and making.

As a means of guidance and delimitation the project has proceeded by establishing a number of productive models, which define significant positions within the map of

the framework. We might think of these as somewhat similar to the strata Deleuze as-

cribes to the methodology of Foucault. They represent pivotal points in the work with a more stable and unitary character than that of the framework as a whole. These relate

to a context that extends beyond the framework itself but to which the framework is

a part. In this they represent an element of knowledge within the thesis relative to a

greater field of experience. Amongst others they act as models for how the framework might approach questions regarding the application of technology, representation and

the nature of the design space. They are naturally developed within a writing practice, but they do not represent the entirety of this practice within the project. Essentially this is because I position these productive models within the methodological structure

as connections between the overall modes of discourse and making. As such they establish verticality between these methodological layers, rather than synthesis and

coherence horizontally across them. The general purpose is to provide a productive

framing for the practice suggested by the thesis; where it can find its specific bearing,

and where the artefacts it produces may point to a precise position within the research. As such they are specific mechanisms that are both technological and conceptual and

which do not articulate what occurs on the discursive side of the work alone, but include elements and observation from both sides. Their usefulness does not peak within the conceptual but the productive.

These directional models are naturally created by means of appropriating, inventing and developing concepts. Though, because the purpose of these models is to provide

directionality for practice the use of concepts has a specific nature within the research.

51


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

Here the use of a concept unfolds relative to how it operates within the specific context

of the work, as opposed to how it might operate within its own historical context or professional discipline. Thus, although a concept might be discovered through the mode

of discourse it becomes articulated relative to the mode of making, which necessarily develops its meaning. The concept must meet the making with a degree of resistance in

order to become productive. The intention here is certainly not to misuse or purposely

misunderstand existing concepts, but to allow the use of concepts in a way that might not have been inherent in the purpose, or context of the original author. This relates to

the form of the practice of writing within the thesis, which I will return to below. Here

I will argue that a concept necessarily and productively evolves together with its context.

The Concept of Practice

Above I described how the overall division of the work into practices based on both

discourse and making allows the notion of practice to permeate the entire terrain of the research. Thus writing becomes a practice that is just as architectural as the production of representations, artefacts and design tools. The question of practice is also seen, not

only as a method to reach the goal of the research, but as this goal itself. Nevertheless, the concept of practice within the research requires further elaboration regarding two questions; what are the concrete models of practice within the research, and how does this link to the necessary context of research?

With a framework for a design practice as the reference point for the work, the practices describing the execution of the project are conducted with the aim of addressing and

establishing the space that this model represents. In this context, and combined with the notion of the research relating to a locally determined nomad science, the practices can be seen to establish a tactical space. In such a space, with reference to Manuel De-

landa, experimental problems must first be embodied in an intensive assemblage prior to producing a solution (Delanda 2005). Simplified, the concept of the intensive – which was also briefly touched upon in the previous chapter – refers to locally determined

mechanisms behind dynamical and qualitatively driven processes; processes that operate bottom-up as opposed to top-down.

“In learning by doing, or by interacting with and adjusting to materials, machines and models, experimentalists progressively discern what is relevant and what is not in a given experiment” (Delanda 2005, p.177).

52


METHODOLOGY

Knowledge is thus gained through actively engaging with a body of material; in the present case one consisting of texts, technologies, techniques, tools, artefacts and visual representations. Sanford Kwinter states that this approach assigns a heuristic model to

practice that is neither exegetic nor deductive. Instead it is genealogical and cartographic (Kwinter 2001, p.40). It represents a tactical approach that belongs to a non-space, which is that of a “shifting, transitory, and volatile materiality of flux and movement” (Kwinter 200, p.122).

This cartography created through the practices of the research is itself aimed towards

addressing the concept of practice. The outcome of the research establishes a map – constituting a kind of experimental design laboratory – by which a design practice might

navigate, and in a sense this makes practice the knowledge produced by the research. Art

historian and critic James Elkins proposes this as a valid strategy for avoiding, what he

considers, the institutionalised concepts of new knowledge and research (Elkins 2009). This proposal is made in relation to a discussion of the topic of doctoral degrees pursued

at a considerable distance from the context of ‘hard’ sciences (i.e. the arts). Practice as understood by Elkins should not be reduced to denote the material evidence produced

by the practice, but understood as the series of acts leading to its creation. According to

engineer and professor of design research Bruce Archer a research activity carried out through the medium of the practitioner must comply with the rules governing research in general, i.e. as a systematic enquiry directed towards a specific goal that produces

communicable knowledge. Though, in one respect it differs from what is commonly accepted: researching through practice must include the practitioner himself, and in this

way can never be interference/value free, or have a non-judgemental basis (Archer 1995). Consequently, such research can according to Archer not be deemed objective in a strict

sense of the word and therefore its findings might only be generalizable to a limited

degree. I already touched upon this topic earlier in the chapter, in relation to consider-

ing the nature of architectural research. I will posit that the methodological structure

of the thesis accommodates this condition of diminished objectivity. Mainly because it displays the preconceptions of the researcher by stating them openly in the form of the directional components discussed above, as well as throughout the fluctuation between

the discursive and non-discursive elements of the research. Herein lays the importance

of defining the writing of the thesis beyond a mere descriptive functionality, as part of

the inventive process behind the creation of the thesis outcome; where writing becomes part of the practices leading to the production of non-discursive research evidence.

53


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

Practice of Writing

I can now return to the statements I made above regarding the use of conceptual elements as part of the directional models (points of orientation) of the thesis and a

corresponding appropriation of concepts. These matters closely relate to the perception of the practice of writing and its form within the project. In the above I define this practice as part of the inventive process of the project. That is, as also part of the

practice of making, as opposed to only a description or prescription of it. Thus, the practices attached to the two operational modes of the thesis combine. Since I also de-

fine practice as positively contaminated by the practitioner then the practice of writing must be affected by this interference as well. This might however appear problematic in the context of research, even architectural. In order to support this situation the

practice of writing within the thesis assumes a specific form chosen to accommodate the methodological understanding of the thesis.

In ‘The Essay as Form’ Theodor W. Adorno argues that between the writing cultures of

an institutionalized science and art a third form exist. This form is that of the essay,

which neither subscribes to the impersonal distancing of the first nor the idiosyncratic

form of the other. The essay form describes a kind of writing that operates precisely between art and science. It distinguishes itself by discussing ideas while being neither scientific theory nor philosophy, and it acknowledges the presence of personal experi-

ence. I will propose that the essay represents a form of writing that allows us to speak from the same intellectual space as that of architecture, being neither science nor art. “The essay […] does not permit its domain to be prescribed. Instead of achieving

something scientifically, or creating something artistically, the effort of the essay

reflects a childlike freedom that catches fire without scruple, on what others have already done. […] Luck and play are essential to the essay. It does not begin with

Adam and Eve but with what it wants to discuss; it says what is at issue and stops where it feels itself complete – not where nothing is left to say.” (Adorno 1984, p.152)

According to Adorno the form of writing exemplified by the essay allows a certain

freedom in appropriating and using what others have written. For some this might appear as over-interpretation, because a prohibition seems to exist towards going

against the intended meaning of a certain text. In this view understanding becomes

54


METHODOLOGY

restricted to just what a given author wants to say. However, the essay introduces

concepts directly as it receives them; “They gain their precision only through their relation to one another� (Adorno 1984, p.160). For Adorno concepts cannot be understood as

something definable within a space outside the context in which they are currently

used. The essay form acknowledges this and forces the meaning of concepts to develop, with the text, and thus it reflectively grasps the meaning of a concept. In this way this

form of writing establishes an intellectual space of mobility, by freeing concepts from

what Adorno designates as orthodoxy. Consequently, by re-contextualizing concepts they gain new or additional meaning while they simultaneously reveal something con-

cerning their point of departure. Importantly, the point of departure is therefore not simply ignored.

The form of writing observed in the position of Adorno, is a valuable idea in the present endeavor because it allows a practice of writing to become part of a space of

invention. In parallel to this notion, the writing of the thesis has been directed towards

establishing a conceptual space that can be described as playful and re-contextualizing, rather than descriptive, prescriptive or legitimizing. Thus, the purpose and form of the

writing does not seek to exhaustively describe a pre-existing (institutional) context, and to legitimize the findings of the thesis relative to this. Rather, it seeks to create components that serve the purpose of conceptualizing the framework for design. This

certainly reflects back towards, and represents contributions to, an existing context. But, it is not subordinated by it.

In closing this section, it should be noted that I do not implement the form of the essay directly as a series of essayistic texts, because I seek to establish a degree of interface across chapters. Thus, the essay is observed more as method than form, mainly representing an operational model for the practice of writing. However, as stated in

the introductory chapter, I do strive to establish a degree of autonomy within each individual chapter. This serves the purpose of emphasizing the cartographic character of the framework, where elements are not understood as a necessary series of successive arguments. Rather, the intention is to establish a situation, where the body of texts

comprising the thesis may be approached through various routes in order to extract different potentials from within the terrain they constitute.

Practice of Making

The concrete nature of the practice of making should also be addressed. As indicated

55


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

by the introductory chapter this practice departs from the utilization of procedural and generative techniques, which are made available through the use of algorithmic and

computational design approaches. In this way making relates to specific digital skills

and a concept of craft with the computer as a medium. In a discussion of such a craft, professor of architecture and information-design Malcolm McCullough observes that

historically technology has supplied an increased abstraction of working processes, providing successive levels of invention which serve to improve upon the processes by which we work (McCullough 1996).

“Each level forms a layer over the old, rather than casting it aside, as in the stages

of natural growth. This means that even if new abstractions become the most

prominent methods, they do not replace existing activities so much as transform or complete them.” (ibid, p.28)

Thus, the digital crafting of architectural artifacts implies the presence of a broader

context of technologies, techniques and practices to which the digital relates. Also, a craft cannot be defined by its tools or a production of form alone, but is characterized by a probing of a given medium’s capacity (ibid, p.29). In the case of the abstract

medium of the computer, this probing holds the possibility to actively reshape the medium itself (ibid, p.115). In this regard, I observe that what fundamentally defines the

medium of the computer is an immaterial, informational and procedural open-ended organization, as opposed to a structure based on material mechanics. Consequently, a

reshaping of the medium must necessitate that the craftsperson understands processes of code writing, since this is the technique by which this specific medium is altered.

Accordingly I perceive a situation for a practice of making, where digital techniques

exist in continuity and relation to a broader network of processes, tools and techniques, and where these techniques should be understood beyond an instrumental reason, as

part of a reflexive qualitative transformation of the applied medium itself by means of the writing of new computational processes. Experiments and Probes

In order to operate relative to this situation I have throughout the course of the project devised a number of experiments all departing from the coding of bespoke digital

design tools or processes intended for such a use. These are not themselves representa-

56


METHODOLOGY

tives of the final framework for design but part of a series of explorations concerning techniques, ideas and concepts, which have contributed with insights needed in order

to make proposals for this framework. Thus, as in the case of the practice of writing, they have been used to identify elements that might stabilise relative to the present context, so as to form a guiding component for future practice. They have served an

exploratory function, relative to the terrain defined by the problem domain of the thesis, and in this role I understand the experiments to produce probes.

Probe is an ambiguous word. It might signify both a physical device and a purely intel-

lectual endeavour. It appears as an object as well as an activity, i.e. as noun and verb, and most importantly it manifests itself differently every time it becomes attached to a

specific profession. A common factor though is that probes always act as instruments

to answer (research) questions. When an experiment is labelled as a probe it designates it with an exploratory nature, rather than making them instrumental in solving

a known problem (Mattelmäki 2008). In this way a probe is something created and

observed in context. Thus, despite the digital outset of the project, as a consequence of relating to the embodied practice of architecture, this has directed attention towards

the production of material artefacts, where a materially manifest reflection describes a dynamic process engaged through technique.

“The fact that the architectural problem is creatively tied to the particularities

of a given context, whether conceptual or actual, and that its practice is held by the media of its invention creates a material focus for practice led architectural research� (Thomsen Tamke 2009)

We have to understand that architectural practice is always touched by the embodied and the material. Thomsen and Tamke show that material evidence act as research

inquiries by which the concepts, technologies and applications of the architectural project can be tested and evaluated (ibid). In the context of research this material evidence should be considered to contain knowledge (Beim Thomsen 2011).

In order to acknowledge this circumstance, as well as the view of McCullough above, three of the experiments attached to the project assume a scope by which a process of

design is pursued from digital tool-making to physical demonstrator. That is, a process by which a generative process has been transformed into a bespoke design tool capable of informing the production and construction of a large scale model. These experi-

57


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

ments are apertures illuminating a complexity encountered outside the scope of the

purely digital, observing that technique cannot be separated from a probing of a material reality. According to associate professor of architecture Peter Bertram: “In contrast to technology, technique is a probe and complexity is not found within the system but in the

material it investigates.” (Bertram 2011, p.48). Thus, the exploration of techniques in a material context – even those identified by digital media – becomes a means to challenge an institutionalised perception of technology as something attached to closed systems of tools and premeditated processes.

The major experiments, spanning across the domains of both digital tool making and a construction of material artefacts, do not however represent the entirety of probes developed over the course of the project. Other experiments have developed through

the thesis at varying degrees of size and complexity. These are generally placed more

singularly within the territory of the digital. I observe that, with a point of departure in the digital, computational and algorithmic, it would be misleading not to emphasize

the presence of the digital output and context in its own right. Therefore a series of

investigations have been conducted for the purpose of understanding and developing

the techniques required in order to master computational media as a tool for design. These experiments do not receive the same degree of explanation and contemplation as the major experiments, but will appear throughout the chapters, not as illustrations of their arguments, but to manifest the double-sidedness of the methodology between discourse and making.

Outline

I will briefly conclude by summarising the methodological assumptions of the thesis: Delimited by an experimental perspective on architectural practice the work is approached with the intention of setting up a framework for a design practice. This

entity is understood as a productive open-ended field – reminiscent of a map – supporting, describing and enclosing a set of interrelated points-of-orientation and areas of movement. In its capacity to function as a map the framework is a device by which

a design practice may productively navigate. Thus, the framework assumes the form

of both the result of the thesis and a methodological element, because it is the organisational principle behind the findings of the thesis. These findings are the result of an

overall operational model envisioned as two parallel pursuits centred respectively on

58


METHODOLOGY

discourse and making, staging a productive conflict between these two. Within this

conflict a series of guiding components are created. These are the elements populat-

ing the map of the framework and relating it to a larger context of practice. They are specific mechanisms that are both technological and conceptual and which do not articulate what occurs on the discursive side of the work alone, but include elements

and observation from both sides. Their usefulness does not peak within the conceptual but the productive.

The methodological framing of the thesis establishes a space, where the concept of

practice spans across a terrain defined by the conceptual, technological, technical and material. Thus, practice relates to both the process of writing and making, and these

practices are mutually part of the creative and productive processes generating the results of the work. In concert, both writing and making are aimed towards creating

local solidifications within the terrain defined by the present problem domain. This, so

as to create the points of orientation by which the specific framework for design – offered by the thesis – might steer.

59


PART 1



3 | TECHNOLOGY BEYOND OPTIMISATION


TECHNOLOGY BEYOND OPTIMISATION

A framework for design is unavoidably also an insertion into something larger and other than itself. It relates to an outside, representing currents and movements, to

which it must be aligned. To this end, the following chapter supplies a point of interface between the work and a larger context. In this task, it is simultaneously general and specific – general because I intend to relate to a broad cultural and societal outlook, specific because I mean to address this from the perspective of a particular

question inherent within the project. This question relates to understanding how a framework for design practice might relate to, and qualify the use of, technology.

Although emphasis is placed on how this discussion might advance a framework for design, the position developed during the course of this chapter is observant of a double-sidedness to its topic. What I notice is that the application of technology is a question equally important to design practice, as the object of the research, and to the

methodology of the research itself. In this the research must be conducted relative to, or in observation of, the very same technologies that will direct the framework. Thus, while the chapter serves to designate a point of departure defining a distinctive technological direction for a design practice, it also serves to set a context for the produc-

tion of knowledge, based on technology, within the thesis. When I proceed through

this chapter I am therefore interested in an intersection between technology, design and the production of knowledge.

Design, Technology and Optimisation

I will suggest that in a digital age the function of technology easily offers a reductionist appearance of servicing architecture as only a form of applied science. This observation

is supported in different ways within contemporary uses of digital technology. I will propose three modes in relation to which this image appears: through analogy, by selflimitation and through an emphasis on performance.

The first, analogy, may be demonstrated through practices of technology-transfer. Here we should be cautious of the differences existing between origin and target in such an exchange, and of the images that might prevail. An example can be observed

in certain design practices where technology-transfer from shipbuilding is seen to establish analogies between buildings and ships, implying a shared context of environ-

mental influences and functional requirements (see Kolarevic 2003, p.8). The original purposes of these technologies are attached to an instrumental reason aimed towards

63


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

[1]

[2]

[3]

1. Shipbuilding industry | 2. Frank Gehry’s Walt Disney Concert Hall in production, Los Angeles, California | 3. Frei Otto, form found roof structure for Olympic Stadium and park, Munich Germany

64


TECHNOLOGY BEYOND OPTIMISATION

the functional efficiency and performance of a ship’s hull. This ascribes a specific ideal

to the goal of the technology. Such an ideal is easily transferred together with the technology itself; thus, legitimizing technology in a specific way, risking to fail in

acknowledging the fundamental differences between the point of departure and the target for the technology-transfer.

Self-limitation, as a means to conceptualize the application of technology as applied science, can be observed when an architectural practice restricts itself to precisely this

limiting view; that is to say, when technology in a design practice is consciously con-

ceptualised in accordance to an instrumental reason. As an example Gehry Partners utilize technology to “allow design to project through the process of realization in the build

environment” (Shelden Kashyap 2012, p.123). Here design is something interfacing with a technology focused on project execution, but where design can exist separately from technology.

When technology is used to emphasize performance in design it easily also conceptualizes itself as part of an instrumental reason. This can for example be observed in the

use of digital simulations of physical form-finding experiments. This method produces

“optimized structural forms from a direct causal relationship between the spatiality of force flow and the generated form” (Kotnik Weinstock 2012, p.106). This method has been

significantly (digitally) re-implemented in architectural design practice with the intro-

duction of digital parametric modelling, exemplified by a renewed interest in the work of, amongst others, Frei Otto (Hensel Menges Weinstock 2004).

It is certainly not my intention to criticize the ideas of working with technology trans-

fer, interfacing with project execution or material performance. In this context the intention is only to show how these technology-uses easily emphasize an image of

technology as servicing architecture, first and foremost, as a form of applied science. It is this image I would like to challenge. If this view is to hold true the application of technology has little choice but to become reduced to the category of optimisation. I

will suggest that the role of technology is something much larger and more complex than that. Since technology has always pervaded architecture, and even more so con-

temporary digital architecture, any reductionist conception of technology will necessarily also reduce the design space. Conversely, if technology is theorized to extend

beyond instrumental reason, as also partaking in a creative and productive dynamism, the relationship between technology and architecture could be one characterized by an ability to generate a qualitative difference. In the search for such a relationship I aim

65


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

to present an argument against standards of technology emphasizing optimisation, in

order to demonstrate that this can never be its principal concern within architectural design practice.

Preliminary Remarks on Optimisation and Technology

To begin with, the concept of optimisation should of course be clarified. By optimisa-

tion I refer to an act, process or method of making, where a design, structure or object

is made to perform in accordance – or as close as possible – to some fixed ideal. This

ideal should preferably be possible to describe exhaustively, so as to supply a clear aim. In the present context an important distinction to make is between optimisation of this kind, where optimisation is applied as a conscious top-down function, and optimisation as an inherent material process independent from a deliberate external causa-

tion. Such processes are, for instance, the ones found in complex systems exhibiting

emergent properties or, in a slightly more mechanical form, material behaviour under stress. In the case of the top-down version, by being related to a pre-set ideal, optimi-

sation leans towards becoming an end in itself, and herby a principle for justifying the application of technology. It is especially in this case an objection is made.

Regarding the other concept of importance here, technology, a few preliminary remarks

should also be made. Often very wide-ranging it denotes a broad range of knowledge regarding tools and their usage. This distinction between knowledge and tool is important because it allows us to detach a paradigm of knowledge from a tool. The way we think, when we apply a tool, affects what it can do, and how it does it. The question

here is what it means to think of a tool as an instrument of optimisation, and what

other ways of reflection can be linked to the application of a tool? It is by answering such questions that a space for the concrete techniques of the framework can materi-

alize. In addition to this relation to knowledge the architect Ali Rahim suggests that technology is the application of a purely technical or scientific advance to a cultural

context; the effectiveness of which lies in its ability to produce new patterns of behavior and levels of performance (Rahim 2006). This provides a need to think of technology relative to the notions of meaning and understanding, and in effect this requires a picture of the cultural conditions to which a technology relates.

In the following I will argue how a paradigm of optimisation – as the representative of an applied science – might prevent technology from effectuating genuine progress

66


TECHNOLOGY BEYOND OPTIMISATION

in architectural design. Here technology becomes constrained to a domain where its task is to restrictively control the performance of an already determined form. The

relevance of questioning this situation lies in the viewpoint that society/culture might give inopportune preference to performance increasing activities, at the expense of novelty, and that optimisation reduces complexity, at the expense of discovery.

Two Models of Society

My reservation towards the concept of optimisation will in the following be related

to two accounts regarding certain cultural and societal conditions, which I link to an understanding of how technology might by appropriated and applied in a design practice. In this role these accounts have been of use in developing the framework, as well

as the research in which it is founded. Each discussion offers an illuminative model for

society and culture by which to construct a point of interface between the work of the thesis and a larger context. The first model exemplifies the mechanisms that support the emphasis on optimisation. The second model offers a perception of society that

exemplifies the mechanisms which might destabilize the processes at work in the first. The first model develops from a critical reflection on how tendencies, within post-

modern culture and post-industrial society, are seen to result in a paradigm of optimal performance. Here the application of technology merges with the production

of knowledge, and in concert both find their mode of justification in a capacity to optimize prevailing institutions and systems. The point of departure for this account is a number of observations made by Jean-Francois Lyotard in ‘The Postmodern Condi-

tion’ (1979) seconded by corresponding observations from Paul Feyerabend’s ‘Against

Method’(1975). Both text sources are first and foremost concerned with the production of knowledge and research. By introducing them relative to questions of archi-

tectural design they are therefore not to be thought of as historical documents in this relation (although they do document a historical situation) instead they act as methodological tools for developing the position of the thesis. As such they are read slightly

outside of the original context of their arguments. What I aim to demonstrate with these references is how post-modern/industrial society created a situation where the

legitimization of knowledge problematically occurs through a paradigm of optimal

performance tied to a specific technological perspective. This perspective can be seen to have relevance for both the framework developed throughout the thesis, and the

67


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

thesis as a production of knowledge by means of technology.

The second model of society is introduced to show that an emergent solution to the

tendencies described by Lyotard might exist in the very structure of contemporary culture. This model departs from a number of observations in recent writings by professor of sociology and cultural studies Scott Lash. Here we find Lyotard’s postmod-

ern culture transformed into a second modernity intensive culture, where an algorithmic conceptualisation of thinking allows a locally determined unfolding of the world. It

is a model where we do not need a dichotomy between a global system and local acts of potential destabilisation. Instead society is described as a vitalist self-organising

system consisting of parallel tendencies originating from the locally determined and the globalizing simultaneously. The key feature within this model, with regards to the

thesis, lies in its positioning of technology outside the realm of science, i.e. instrumental reason, allowing it to interact with reality as a catalyst of difference.

When I choose to introduce both models here it is because manifestations of both can

be observed in a contemporary digital practice of architecture. Professor of architectural history and technology Antoine Picon notes that:

“[…] the only way for architecture to escape the role of legitimation that was as-

signed to it by ruling powers was to be reflexive and critical. By the same token there was an implicit link between the computational and the critical perspectives. Such a link was not without contradictions, for the computational dimension in architecture pointed both at a purely performative attitude, like the one

envisioned by Lyotard, and to its critical counterpart for which performance

mattered less than reflexivity. How was one to reconcile these two seemingly irreconcilable objectives? Contemporary digital architecture is still confronted with this question.” (Picon 2010, p.47) Societal Model I: Homeostat

In an endeavour to account for the status of knowledge-production in a post-modern culture and post-industrial society Jean-Francois Lyotard observes that society assumes a homeostatic character during the second half of the 20th century. It has be-

come a closed functional whole with no need for unifying and universal narratives such as those of the enlightenment or Marxist emancipation (Lyotard 1984). Lyotard observes that this fundamental shift occurs simultaneously with the introduction of

68


TECHNOLOGY BEYOND OPTIMISATION

[4]

4. Ross Ashby’s Homeostat

69


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

new technologies related to computation, which he clearly associates with the field

of cybernetics. This field of enquiry is implicitly criticized throughout the argument. However, I will argue that this criticism misses the self-doubt that occurred within

cybernetics itself, which led to the notion of its so called 2nd order form (Heylighen

Joslyn 2001). Here the mechanistic and prescriptive nature that Lyotard assigns to

cybernetics gives way to a more dynamic and reflexive equivalent. With this in mind, I read Lyotards critique as directed towards first order cybernetics solely; and as such supplementary to the self-critique excised by this field itself. For this reason I characterize the societal model which Lyotard describes by means of the image of the

Homeostat, the self-stabilizing device invented by the 1st generation cybernetician

Ross Ashby (Ashby 1960).

The (cybernetic) technologies defining the post-modern culture and post-industrial

society is by Lyotard observed not to legitimize knowledge in terms of being ‘good’ or ‘just’, but rather through a notion of optimal performance. As a result a prevailing view

of society as a self-regulating mechanistic system takes hold. The goal of which is to optimize the global relationship between input and output:

“Even when its rules are in the process of changing and innovations are occurring, even when its dysfunctions inspire hope and lead to belief in an alterna-

tive, even then what is actually taking place is only an internal readjustment, and its result can be no more than an increase in the system’s ‘viability’. The

only alternative to this kind of performance improvement is entropy, or decline.“ (Lyotard 1984, pp.11-12)

This simplification of society into a system of input and output brings about a problematic tendency to isolate a positivist technological thinking from a critical qualitative. Paul Feyerabend disapprovingly notes that, historically, the production of knowledge has in this way been simplified by simplifying its participants. Conventional

science restricts areas of research within fixed definitions, and hereby separates indi-

vidual domains from other areas of inquiry. This eventually uniforms actions within this domain. Here the ability to create ‘stable’ facts occurs from inhibiting intuitions

that might lead to a blurring of boundaries. This is not desirable because the world

is a largely unknown entity, which asks of us to keep our options open (Feyerabend 2010). Similarly I will propose that we should consider if isolating the application of

70


TECHNOLOGY BEYOND OPTIMISATION

technology within single domains inhibits architectural design’s ability to contribute with discoveries. In my reading of a digitally produced architecture, in this way it often

seems to exhibit a tendency to position technology either within a kind of positivist (e.g. building information modelling) or a formalist speculative trajectory; which, as will be elaborated in chapter six, both assume a great deal of determinism on the part of the architectural model.

Lyotard observes that the self-regulating nature of a performance driven society pres-

ents us with a paradox in relation to the question of discovery. He observes that the presence of novelty and unexpected actions are in fact operational as a means to in-

crease the performativity that the system demands and consumes. At the same time

the nature of any system based on performance maximization must have its goals prescribed in advance, which leaves little space for the unexpected. It simply becomes a

disturbance to be eliminated, disregarding the fact that a qualitative difference within

the system could have been made. What we are left with is a system of countermoves without the ability to produce difference (Lyotard 1984).

Design is inherently concerned with seeking out novelty. That is, we design to make discoveries and to add something to the world which was not there before. In this, the

production of knowledge and the practice of design seem to have coinciding goals, although purposes might differ. Within a society perceived as a self-regulating system the goal of design as well as research must then be to displace and disorient the system

in order to provoke an unexpected reaction; assuming a critical method against repetition. Feyerabend hold that methodology, strictly speaking, is inherently incapable

of replicating the complexity of the errors, prejudice, passion and even questions of style that throughout history has resulted in novelty. A maze of interactions describes

the history of discovery that amongst others contains ideas, interpretations of facts, problems created by conflicting interpretations and mistakes. The context of discovery necessitates an opposition to the dictates of an institutionalized reason, and history

shows that all rules, however plausible, and however firmly grounded in observation, are violated at some time or other. In fact established forms of reason depend on

the fact that older forms of reason were violated at some point or other. These viola-

tions are not accidents, but necessary events of progress, and are both reasonable and

absolutely necessary for the growth of knowledge (Feyerabend 2010, e.g. see p.115). The production of knowledge is in this way a continuous process, where simplicity, elegance and consistency are never necessary conditions of practice – scientific or oth-

71


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

erwise. In this way Feyerabend and Lyotard shares the need to emphasize the importance of introducing disturbing elements within the context of practice.

Although critical and positive thinking tend to separate within the postmodern con-

dition, the classical borders between other areas of inquiry do start to dissolve. Dis-

ciplines disappear or overlap, while new territories are being born, breaking with the historical isolationism of the sciences identified by Feyerabend. As a result the speculative hierarchy of learning is exchanged by an immanent and flat network of areas of

inquiry, the respective frontiers of which are in constant movement. The principle of a universal meta-language is replaced by the principle of a plurality of locally defined systems capable of arguing the ‘truth’. In order to produce proof within this plurality

of argumentative systems the result of an experiment needs to be observable, because it does not any longer naturally fit into a universal paradigm of legitimation.

This is where technology once again enters the discussion seen by Lyotard – with clear

reference to cybernetics – as prosthetic aids for the human organs or as physiologi-

cal systems, whose function it is to receive data or condition the context. They follow a principle of maximizing output (the information or modifications obtained) and minimizing input (the energy expended in the process). Technology is therefore not a

method aimed towards the true, just, or beautiful, etc., but towards efficiency. Here a

technical action is “good” when it does better and/or expends less energy than another. The production of proof is in this way linked to a principle of legitimation where the

goal is an improved performativity; defined as the best possible input/output equation. This means an abandoning of the humanist, as well as idealist paradigms of legitima-

tion (Lyotard 1984, p.46). For Lyotard the consequence hereof becomes a subordina-

tion of quality under efficiency. However, as I will discuss in chapter 5, a cybernetic figure like Gordon Pask, arrive at a different conclusion. Cybernetics and technology

might very well imply a challenge to humanism and idealism, but does not necessitate the use of reductionist input/output equations. For Gordon Pask what materializes

between technological devices and the individuals engaging them is an aesthetically potent environment in which precisely qualitative differences can emerge by a decentred agency shared by man and machine.

Lyotard’s error, or simplification of cybernetics, is nevertheless not what is of importance here. It is the appearance and reading of technology within a specific model of

culture and society. It is not its theoretical point of origin but the observation of its actual manifestation contaminated by all the relations in which it participates. Amongst

72


TECHNOLOGY BEYOND OPTIMISATION

others, what is noteworthy in the account of Lyotard is that optimisation, by means of technology, is shown to assume a principle of determinism. In order to accommodate this assumption a reduction in complexity acts as a prerequisite to action, but by

exchanging complexity with predictability we also reduce the likelihood of novelty and difference. What is observed is that potentials for change are unveiled only as we transcend trivial input/output relations by approaching the world as it is in fact, para-

doxically, theorized by postmodern science itself; as catastrophic, fractal and complex (Lyotard 1984, p.60). Lyotard’s solution to this is to alter our methodologies by seeking a principle by which knowledge can be produced within a narrative understanding

of knowledge, described as a plurality of small stories that function well within the particular contexts in which they apply. This principle, which he calls paralogy1, is to

point out presuppositions in a field of knowledge and to request agents within it to

accept different ones. In essence, paralogy is the act of making a novel move against the emergence of unifying theories, within a plurality of locally defined narratives. It is necessary to hypothesize the existence of a power that destabilizes the capacity for explanation. That is, not without rules but locally determined. This should according

to Lyotard make discoveries unpredictable; where new knowledge becomes the result of catastrophic events (Lyotard 1984, p.60).

In the above model we are presented with the mechanisms that support an empha-

sis on optimisation. Society/culture is described as a closed functional whole, which, rather than being driven by ideals, defines its goal in terms of continuing improvement upon its own performance. Consequently, it maintains already defined institutions and

pacifies a critical qualitative thinking. However, criticality is shown to be the necessity required in order to produce another condition, in which difference and novelty is possible. With regards to the technological nature of the thesis, critical and technological

thinking are within this model separated. Consequently, computation as understood within this model points towards the very mechanisms that are the problem. Compu-

tation is at the service of the continuous readjustment of a system to pre-determined goals. However, as Picon quoted earlier observes, computation in contemporary digital architecture is as well part of critical movement. This suggests that it might be possible

to address questions of criticality and technology differently, and with a more contem1. Literally refers to a move against logic or reason, but should here be understood as an action made against conventional and repetitive reason.

73


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

porary outlook than that of Lyotard. Or, since the problems he addresses can still be observed, that a parallel, alternative model exists as well. Intermission: Criticality, Resistance and Alternative

The criticality of Lyotard – and Feyerabend for that matter – does possess a risk of

becoming a kind of negatively defined position, thus being shaped by the very thing it

opposes. It is resistance to modernism, as well as the condition that followed. Political philosophers Michael Hardt and Antonio Negri observes that postmodernity marks

the end of the core elements of modernity but is “[…] primarily a negative designation, focusing on what has ended.” And does not “capture a strong notion of resistance or articu-

late what constitutes “beyond” modernity” (Hardt Negri 2009, p.114). There is a general

limitation to the concepts and practices of resistance, in that they easily get stuck in an oppositional stance. A move from resistance to alternative is required. In relation to this Hardt and Negri propose the concept of altermodernity:

“Altermodernity has a diagonal relationship with modernity. It marks conflict with modernity’s hierarchies as much as does antimodernity but orients the forces

of resistance more clearly toward an autonomous terrain.” (Hardt Negri 2009, p.102).

Altermodern criticality indicates a break with modernity and the power relations that

define it by evolving through traditions of anti-modernity. However, it simultane-

ously opposes these traditions by extending beyond mere opposition and resistance. It breaks with anti-modernity, refusing dialectical opposition, and moves from resistance

to the proposition of alternatives. It possesses a transformative mechanism by which

opposition is freed from being negatively defined. Such transformation occurs within a model reminiscent of Lyotard’s plurality of small narratives; a horizontal network of

the singular and autonomous (Hardt Negri 2009, p.112). The figure of the intellectual is here described as “[…] a militant, engaged as a singularity amongst others embarked on the project of co-research aimed at making the multitude” (Hardt Negri 2009, p.118).

Lyotard’s position does hint towards an alternative, but I will argue that it does not

reach a point where it describes an altermodern condition per se. Paralogy is a mechanism within the minor narratives, a non-modern horizontal network, but it is only a mechanism, or vector, by which change might occur.

74


TECHNOLOGY BEYOND OPTIMISATION

Lyotard’s position also exemplifies problems attached to the negatively defined in

its opposition to technology. Because technology is part of the problem observed by

Lyotard, his post-modern resistance also assumes a form of opposition to technology. But what if an altermodern vector in fact frees technology from the problems of the

homeostatic model? So that technology is no-longer at its service. This might explain

the presence of technology (in the form of computation) in critical as well ‘conformist’ practice.

Societal Model II: Algorithmic

What follows can be conceptualised along the lines of precisely freeing technology

from being at the service of a dominant power. Without claiming equivalence, something resembling the principle of paralogy is now in fact observable within today’s culture, however, it has become technologized.

Scott Lash observes that rules and reason – within what he identifies as a contemporary mediatized second modernity – have ceased to be regulative. Instead they have

become generative and informational, and as such algorithmic (Lash 2007, p.72). An

algorithm is always determined by its unfolding from a local point within its environ-

ment. At the level of mechanisms this relates the concept of algorithm to the function of paralogy. However, the society described by Lash is different from that of Lyotard in two important ways.

Firstly, culture is not theorized as a dichotomy between a global self-regulating system

and locally determined acts of potential destabilization. Instead culture is according to Lash characterized by two simultaneous and interlinked tendencies; an intensive and

an extensive (Lash 2010, pp.1-5). The latter is commonly experienced through global-

ization, where capitalism, culture, society and politics become increasingly extensive. This means a universalization of contemporary culture characterized by geographical spread and homogenization; a culture of the same. Parallel to this tendency, we also

observe the phenomenon of the intensive. Intensive culture is a culture of difference; where units – anything from individuals to brands – assume heterogeneity. The impor-

tant observation in this regard is that, although seemingly opposed, it is this intensive culture that gives birth to the extensive. It is the intensive that generates what we encounter, and as such an intensity is an algorithm producing extensities. This means

that extensities are fixed while intensities are always in process. Extensities are ‘beings’, and as such in equilibrium, while intensities are ‘becomings’. They represent a move-

75


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

[5]

5. Algorithmic network/branching morphology, by author. From the ‘Non-Linear Systems Dynamics’ workshop at Smart Geometry 2010

76


TECHNOLOGY BEYOND OPTIMISATION

ment, not in space, but a movement of change and instability in the intensity itself.

The second difference between the positions of Lash and Lyotard relates to the concept of self-organization. Lash also maintains that the organizational principle of society occurs from the inside-out as self-organization. This is inevitable as long as society is

theorized as a complex system, given that self-organization is a general property of such systems (Cilliers 1998, p.90). However self-organization now assumes a different

character. It does no longer denote the mechanistic principles of 1st order cybernet-

ics, but the dynamic and reciprocal autopoiesis that followed (see Heylighen Joslyn

2001, and chapter 5). Thus, we are dealing with evolving systems without exteriority in which observation includes participating as a part of the system itself.

Within a society theorized as self-organizing and intensive the social itself becomes, on the one hand, molecular and algorithmic, and on the other, globalizing. In both

cases it disrupts geographically defined regulative norms. Individuals become rule finders rather than rule followers, and as such reflexive (Lash 2007, p.82). What is important to note here is that intensive culture can be perceived as a critique of instru-

mental reason. We can approach an object either as extensity or intensity; from the surface of interface or the depth of algorithmic generation. In the case of the latter we

approach the ting virtually, as singularity and not by means of fixed categories, while

to merely know a thing extensively in terms of categories is to know it instrumentally. This is because categories are only convenient abstractions that have no place in inten-

sive thinking, since this can be characterized as populational thinking (Delanda 2005, pp.58-59). Populational thinking eliminates the need for pre-existing archetypes and archetypes as ideal norms. Forms do in this way not pre-exist the population they are

rather statistical results. Obviously the conceptual language used here is transferred

from the realm of biology. As an explanatory analogy from this domain the following quote from D’Arcy Thompson might clarify what is meant: “(…) in the representation of form and in the comparison of kindred forms, we see in the one case a diagram of forces in

equilibrium, and in the other case we discern the magnitude and the direction of the forces

which have sufficed to convert the one form into the other” (Thomson 2010, p.270). To know something intensively is to break with instrumental knowledge.

Intensive culture is related to life, it is a vitalist system as opposed to a mechanistic

(Lash 2010, pp.21-23). In the case of the latter cause originates from an outside as a unidirectional input/output situation. Vitalist systems, on the other hand, are self-

causing. Energy is internally generated whereby causality proceeds from the inside.

77


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

These systems are characterized by becoming, deterritorialization and non-linearity. In addition, another distinction between mechanistic and vitalist systems is in how

time is conceptualized. Mechanistic time is Newtonian time, and therefore reversible (Prigogine 1980), divided into discrete and interchangeable units, where memory has no role and no life. Vitalist time, on the other hand, is irreversible, continuous and seamless. This kind of time is always attuned to goals as internally generated value and dependant on memory.

Approaching reality as a vitalist system fundamentally alters how we interfere with our surroundings. Lash observes that within the 1st modernity nature was a target of

rationalization. It was observed as a mechanism by a reason external to it. Here reason takes nature for its object. The assumptions of rationalized nature are atomistic: of identical parts making up wholes, where one can take the place of the other without

change. In this, nature loses agency and life is drained from it. The mediatisation of contemporary culture changes this:

“Matter – like genetic matter – that reads and stores and transports is intel-

ligent. Reason enters matter, as it is mediatized. Now reason is no longer just outside of matter. Intelligence is no longer just outside. Intelligence is distributed. Media presume distributed intelligence. Media presume a certain dying

of the author. Rationality or reason, once it is distributed, becomes reflexivity. Thus nature becomes intelligent, becomes pervaded by reason as it is mediatized.” (Lash 2007, p.74)

As nature becomes intelligent, in view of the fact that it becomes artificial and cyber-

netic, it becomes technological as well. Therefore it is no longer solely the object of science, since it is important to note that science is different from technology. Science

finds and discovers, and at least classically, presumes that reason is outside nature.

Lash observes that when reason is positioned here it can only pose ‘what’ questions, while reason inside nature asks ‘how’ questions. These questions are like instructions, algorithms, i.e. a technological intelligence (Lash 2007, pp.74-75).

Information/media society involves a technologization of areas of science. Biology becomes biotechnology, mathematics becomes computer science, and in this, science is partially transformed into engineering.

78


TECHNOLOGY BEYOND OPTIMISATION

“There is a very strong engineering principle or logic in mediatization. Science

and art, which in the first modernity took their distance from engineering, now find themselves pervaded by it.” (Lash 2007, p.75)

Questions of optimisation and performance of course closely tie to engineering. In

this, a strong contemporary tendency towards optimisation again seems logical. But

optimisation, through its reliance on determinism, depends on the ’what’ of science, whereby reason, and agency, is placed outside of the object. If technology is about ‘the how’ it needs to address reason inside the object.

What Lash presents us with is a paradigm of technology, where nature and culture

is collapsed into one shared artificial field. Here technology always exist as part of an

intensive practice that unfolds as bottom-up process. In this it makes top-down optimisation highly difficult because practice never occurs from an outside. The question

is what kind of realization that takes place here. A vast field of knowledge opens up and it is essential that we do not fall victim to conventional mode of thinking should we address this potential.

An approach to Technology

The aim of this chapter has been to develop an understanding of the broad cultural/

societal context to which the thesis relates; this with implications for the project as both framework for design and research with a focus on questions of technology.

Both models presented assume the form of self-organizing systems without exterior-

ity, or an overarching defining unity other than that of the system itself. However, each model presents us with a choice for how technology might be understood and

approached. I will hypothesise that both choices are possible today, and that it is necessary to define research and design concerned with the use of technology in relation to these.

In the first model technology is at the service of the system in its prevailing con-

figuration and form. One might say that technology is applied only to its extensive structure. In the second model technology is at the service of the qualitative forces at

work within culture and society. In the first model society tends to legitimize the use

of technology by reference to an ability to increase performativity. In the second model technology finds legitimacy by designating how one might productively draw out the

79


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

autonomous agency of things. In the first model technology is utilized in order to produce stability, while in the second model it is used to maintain a constant of change.

Both models observe that the conditions for discovery and change reside in the locally determined. This means that it is not within the unities or extensities that practices

of design or research find their bearing in a potential alter-modern diagonal. That is, in a movement away from something in the form of an alternative, rather than in a normative opposition.

I believe that Lyotard and Fyerabend can be used to show that if we simplify technol-

ogy to solely concern increased performativity we isolate design from the qualitative aspects that – in concert with the technological – might lead to good and new design

solutions. If we wish to continue the application of new technologies in architecture, we therefore need to question how we legitimize this application. A principle of legitimization based on an ability to produce qualitative difference and novelty might

be a candidate. This would require that our techniques are aimed at discovery, seeking out the productive accidents that lead this way. Optimisation is opposed to such an

approach because it resists disturbances and the blurring of boundaries. Its legitimation occurs at the expense of qualities, and its determinism prevents novelty. In a sense

‘good’ architecture is not optimal architecture because it needs to relate to a model of reality that is too complex to be presented though an absolute representation.

With regard to the second model advocated by Lash, it represents a culture that is describable through many of the same concepts as contemplated by the thesis. The

intensive/extensive culture is reflexive rather than representational, generative rather

than planned. This suggests that the framework for design developed through the

course of these pages resonates with certain tendencies already at work in a broader context. In terms of how the model can be used more concretely, I appropriate Lash’s

vitalist perception of technology. It represents a required critique of the instrumental reason at work in Lyotard’s model; by amongst others arguing that technology is dif-

ferent from science. By approaching reality as a vitalist system, nature can no longer be

a target for rationalization, where reason exists external to it. Nature has agency and

this is something I desire to maintain within a framework for design. Throughout the rest of the thesis it is this vitalist approach to technology, beyond optimisation, that will be the guide.

80


TECHNOLOGY BEYOND OPTIMISATION

[6]

[7]

6. & 7. Images of a model resulting from the ‘Non-Linear Systems Dynamics’ workshop at Smart Geometry 2010 (as featured in the exhibition ‘Working Prototypes’ at DHUB, a contemporary design museum in Barcelona, Spain). The workshop was led by Jenny Sabin, Peter Lloyd Jones, Andrew Lucia, Erica Savig, with the author as participant and contributor. The project explores a condition of technology transfer between architecture and systems biology as a means to gain insight into dynamic living systems for the development of novel computational design tools and material systems. Generative design techniques emerge with references to natural systems, not as mimicry but as trans-disciplinary translation of flexibility, adaptation and complexity into realms of architectural manifestation.

81


4 | A MANIFOLD DIGITAL


A MANIFOLD DIGITAL

In the following I will consider the specific context of a digitally conceived architecture to which the research and its proposed framework for design relates. The purpose of this investigation is to position the work relative to this context, indicate relevant

discourses within it and to point towards where the work itself, reflexively, contributes to the field. The findings of the previous chapter are used to designate my approach in

this task. Here the positions examined earlier can be seen to propose certain require-

ments and restrictions when making a contribution to knowledge. These constraints will be used to assign a specific reading of the context, as well as the position assumed

within it. As such, my intention is not a historic or exhaustive account of the field of digital architecture. Rather, the goal is to clarify how the framework aligns to adjacent

practices and tendencies of a digital architecture, and as a result synthesizes a productive position from within this field.

Since the findings of the previous chapter are important to my approach, I will shortly

summaries the key points that will be impose in this chapter, and do so relative to the

present change in perspective. I will once again divide the argument up relative to the

two societal/cultural models, since their respectively attached positions assume different roles in the present endeavour.

In the first model, attached to the concept of homeostasis, technology was seen to be part of an instrumental reason operating independently from a critical qualitative thinking. The notion of optimization was problematized as the standard by which

the use of technology found its legitimacy. Thus the application of technology assumed the form of an applied science at the service of improving the performance of

predefined goals within a system (society). Amongst others, this results in a problem

of restricting thinking to specific stable institutions, with clearly defined boundaries. When considering digital practice in architecture, relative to these observations, my

interests lie in elements and tendencies within this field, which do not fall victim to becoming instrumental to a homeostatic model of society. I will suggest that this requires two things. Firstly, technical knowledge will need to be combined with a critical

qualitative knowledge in order to avoid being limited to an instrumental reason. A technical knowledge is necessarily to be considered important to a digital practice in

architecture but not in isolation. Secondly, this criticality should not seek the ideal of unifying structures, which must be considered in opposition to a productively conflicting project.

This perspective has consequences for the present investigation. The emphasis on a

83


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

qualitative critical thinking entail that I, in the following, will disregard positions where the digital nature of a practice is separate from this kind of reflection. Here I am

specifically referring to the idea of digital technology as an extension of existing tools and practices; a situation where technology is essentially superimposed onto a predigital design process, for the purpose of increased efficiency. This is by Kostas Terzidis defined as a situation of computerisation, as opposed to one of computation (Terzidis

2006, p.xi). What is of interest is how digital technology is applied critically. Though, as implied by Hardt and Negri in the previous chapter criticality can nevertheless as-

sume a form where it becomes defined by what it attempts to abandon – consequently not contributing with genuine change. In this, a move away from resistance towards

alternative and autonomy is required. It is these aspects within digital architecture that I will focus on.

In the second model, attached to the concept of the algorithmic and generative, tech-

nology is seen to be part of a liberating reason departing from a locally determined and qualitatively driven position. In this model we encounter destabilizing intensive processes continuously producing heterogeneity and difference in the presence of uni-

fied, or unifying, structures. Regulative rules and reason are challenged by counterparts of a generative and algorithmic nature. When related to the observation of digital practices in architecture the second model is of interest because it methodologically

offers a useful mode of reading their tendencies and contributions. What the model

suggests is that tendencies within digital architectural practice can be interpreted intensively, as singularities, something defining on-going processes. In this way I aim to

describe how different practices can be used to instantiate different processes within the proposed framework for design and potentially beyond.

From another perspective the second model can also be seen to coincide with theo-

retical tendencies directly present in many practices of digital architecture today. Here I refer to the intellectual currents characterized by concepts related to a becoming

of form, inspired by findings in contemporary natural science and philosophy, i.e. a generative perspective. This apparent intersection between contemporary cultural conditions, theory and digital design leads me to accentuate precisely trends in digital practice that exhibit these shared tendencies.

In the following I will present two discussion related to a point of departure in the above; one concerning the overall field of digital architecture, and one concerning concrete tendencies within this field. Firstly, I will investigate the overall nature of

84


A MANIFOLD DIGITAL

a critical digital architecture by addressing a central question: Can the contemporary field of digital architecture be perceived as a unity? This would complicate the intentions expressed above, by restricting architectural thinking to an established construct with

visibly defined boundaries. Here the context of digital architecture would have to be

addressed as a current grand narrative for the profession. By contesting that this is the case I proceed to the second discussion, which depends on the opposite situation, that a condition of unity does not define digital practices. Here I will discuss a number of

evolving tendencies, by which to relate the project. These will in accordance to the guidelines sketched out in this introduction be investigated according to how they

assume criticality, represent difference and produce diversification as singularities of on-going processes. Processes that I seek to align the framework to.

(The Question of) Unity

Recently it has been suggested by Patrik Schumacher – a partner at Zaha Hadid

Architects and researcher at the Architectural Association School of Architecture in

London – that the avant-garde of a digitally designed architecture has in fact assumed

the guise of a new unified style, unparalleled since modernism (Schumacher 2011, 2012). Schumacher terms this style Parametricism and insists that this project incorpo-

rates all tendencies followed by a digitally informed (progressive) architecture over the course of its short history. Thus, if he is correct, Parametricism equals the architectural

context of the thesis, and would as such assume a prescriptive relation to it. For this

reason I reference the position of Schumacher here, even though, as will be discussed, I do not subscribe to it.

When Schumacher uses the notion of Style it is to be understood in the sense of what he calls a design research program (Schumacher 2012, p.643), and as such defined

beyond a mere notion of fashion and appearances. A design research program should according to Schumacher be conceived similarly to how paradigms frame scientific

research programs. Thus, a new style in architecture equals a new paradigm in science, as something that redefines the fundamental categories, purposes and methods of a

coherent collective endeavour. In Schumacher’s view this orientation towards the im-

age of a scientific practice makes the principles of Parametricism unbreakable because they become necessary conditions to test the style’s claim of universal competency. Its alleged performative advantage over all previous styles is premised on a consistency of adhering to these principles. Thus, Schumacher subscribes to exactly what Feyerabend

85


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

[1]

[2]

[3]

1. Schumacher’s parametricism. Abu Dahbi Performing Arts Centre by Zaha Hadid Architects (2008) | 2. Early project by Greg Lynn, Cardiff Bay Opera House Competition, plan (1994) | 3. Greg Lynn, Model, Embryological House (1998-1999)

86


A MANIFOLD DIGITAL

in chapter three is shown to oppose in a conventional view of science, because, according to him, it restricts the possibility of discovery and change.

As a style Parametricism comprises a series of principles stressing the importance of operations related to fluidity, gradients and variety in both formal and functional

matters. Schumacher relates these principles to the range of tools and techniques

made available by the introduction of computer technology within design practice. This includes the use of topological geometry, time based processes and the simulation of generative and evolutionary processes – and importantly signifies a move away from

a design space defined by an absolute representation towards one that is parameterized and relational.

“Parametricism implies that all architectural elements and complexes are para-

metrically malleable. This implies a fundamental ontological shift within the basic, constituent elements of architecture. Instead of the Classical and Modern reliance on ideal […] geometric figures […] the new primitives of Parametricism are animate […] geometrical entities […].” (Schumacher 2012, p.654)

With these principles, tools and techniques Schumacher positions himself along a

recent historical lineage of a certain architectural discourse. This discourse signifies an important intersection between a philosophical reorientation of architecture and

a new technological terrain by which post-modernism and Deconstructivism transformed into a foundation for a contemporary digital architecture. A significant proponent of this transformation is the architect Gregg Lynn, who in his seminal 1993

essay “The Folded, the Pliant and the Supple” (Lynn 1998) exemplifies this discourse. As an example of a point of departure for the processes leading to the proposed unifying

style of Schumacher, the text has importance for my position regarding the existence of unity.

In the vocabulary presented in chapter three, Lynn represents an altermodern stance by

insisting that neither reactionary modernist unity, nor an avant-garde dismantling of it seems adequate as a model for contemporary architecture and urbanism. Evidence towards another position is observed in the form of architects whose work can be

described as post-contradictory, represented through the formulation of an alternative smoothness. Regarding the sources of this work Lynn writes:

“[…] topological geometry, morphology, morphogenesis, catastrophe theory or

87


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

the computer technology of both the defence and Hollywood film industry - are characteristics of smooth transformation involving the intensive integration of differences within a continuous yet heterogeneous system. Smooth mixtures

are made up of disparate elements which maintain their integrity while being blended within a continuous field of other free elements.”(Lynn 1998, p.110)

This condition of smoothness, in the form of pliant and viscous systems for design, leads to the proposal of a new architectural operation of folding – a means to integrate unrelated elements within a new continuous heterogeneity. The result is an

architecture operating by means of a compressed continuous layering “within which

heterogeneous deposits are still intact in varying degrees of intensity.”(Lynn 1998, p.112) Thus, Lynn subscribes to neither the homogeneity of (neo-) modernism, nor the heterogeneous, fragmented and discontinuous theories of conflict that followed it. The

spaces he propose possess a curvelinear sensibility (drawing inspiration from Deleuze)

“capable of complex deformations in response to programmatic, structural, economic, aesthetic, political and contextual influences.” (Lynn 1998, p.115). As such, folded space

should be understood as more than the folding of form – as also a folding of ‘forces’, external as well as internal.

The project of folding described by Lynn is linked to observations of its manifestation in a range of architectural discourses. In terms of a possible unity between these ‘The Folded, the Pliant and the Supple’ suggests that these discourses share interlinked

elements at a conceptual and tactical level. Conceptually they represent an alternative description of spatial complexity based on topological thinking, and; “They use tactics and strategies that are compliant to, complicated by, and complicit with external forces in

manners that are submissive, suppliant, adaptable, contingent, responsive, fluent, and yielding through involvement and incorporation.“ (Lynn 1998, p.130)

It is Schumacher’s view that what followed this conceptual and tactical project was more than a new set of techniques – it was a new unifying architectural movement: “The techniques in question – the employment of animation, simulation and form-finding tools, as well as parametric modelling and scripting – have in-

spired a new collective movement with radically new ambitions and values” (Schumacher 2012, p.641)

88


A MANIFOLD DIGITAL

Within an area of collective design research emerging from within this field Schumacher believes his principles have developed to the point where they can now be ex-

plicitly stated to have become a priori premises of a new generation of architects. Schumacher therefore defines Parametricism as a mature style, where the shared concepts, computational techniques, formal repertoires and tectonic logic that character-

ize work such as described by Lynn are crystalizing into a solid new paradigm for architecture. Thus, he insists that a unified style can exist in the 21st century, although this might be considered an anachronism by many theorists assuming that Modern-

ism would be followed by an era of stylistic pluralism. Schumacher is against pluralism

of styles because he believes that it results in a condition of stylistic fragmentation that disrupts the Parametricist search for continuity and intensity of relations.

I will now address my own position in these matters. Schumacher consolidates the

tendencies by which a critical digital architecture operates into a unity, and in this, makes the field prescriptive. He attempts to make this characteristic unproblematic by

stating that the field has ‘matured’; consequently, making any further development of a digital architecture internal to itself. In this the project can be said to assume the form

of extensity or a ‘royal science’. This leads Schumacher to appropriate a conventional

scientific rigor as an ideal for methodology. By placing his project in direct linearity with projects such as Lynn’s, Schumacher fails to observe that the project described by Lynn follows the diagonal of the alternative and is not merely resistance to what came

before. Schumacher’s project solidifies as resistance to modernism as well as postmodernism. Neither does it correspond to the image Lash conveys of contemporary

culture in its use of regulatory rules. It is simply extensively defined, thus the inten-

sively defined forms it advocates become detached from the conceptual framework to which they comply, and so, become imagery.

Although I am interested in the consolidation of tendencies within digital practice as well, I subscribe to the double movement of the intensive and the extensive – the

nomad and the royal sciences. Here unities are preliminary unstable constructions, which can and should be challenged, if the goal is change. In this view reality consists

of things that are separating and things that unite. For example professor of architecture Branko Kolarevic holds that although: “[…] no monolithic movement exists among

the digital-avant-garde in architecture. What unites digital architects, designers and thinkers is […] the use of digital technology as an enabling apparatus that directly integrates

conception and production in ways that are unprecedented since the medieval times of master

89


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

builders� (Kolarevic 2003, p.4). This suggests an alternate route to unity. While Schumacher solidifies the tendencies of a digital architecture as a converging unity, we might instead think of unity as something occurring momentarily between tendencies, where these tendencies are not slowing down themselves.

It is in this way that I perceive the framework of the thesis; a position within a system of tendencies, by which it is affected relative to distance and proximity. As such the

framework is not prescriptive but suggestive. Indicating how we might tap into cur-

rent trends within a much greater dynamic field of digital progress. It is not an attempt to unify the field, but a momentary solidification within it.

Tendencies within the Field

Consequently, and in contrast to Schumacher, I adopt the position that the field of critical digital architecture is still maturing and has not solidified in the form of an excluding and inert unity. Instead I suggest that critical digital architecture as a whole

does not represent an architectural movement, but can be conceptualized as a shared

field defined by certain contemporary tendencies, working their effect, in different ways, within independent architectural practices. As stated in the introduction to this

chapter, as an explanatory image, I propose to understand these tendencies as intensities defining an on-going evolutionary process of architectural design. As such the

tendencies are constructions that do not exist concretely in any practice, but define singularities around which concrete practices emerge. I observe this to manifests itself in the form of blurred boundaries, parallels and overlaps between the focuses of individual practices, which is not to be confused with unity but simply is a result of occupying the same contemporary landscape.

In developing the framework for design, suggested by the thesis, I identify four such overarching tendencies by which I draw relevance. They represent areas in which their

proponents observe potentials attached to the use of digital technology, by which a

contemporary technological landscape achieves applicability and legitimacy. By means of introduction, and slightly simplified, I suggest naming these tendencies by what

they represent in terms of technique, focus and principles in a digital practice; the generative, the material, design realization and the cultural reflexive.

1. The generative tendency is of course the pivotal trajectory of the thesis. Observed in the broad context of digital architecture it assumes two forms; as technique and as

conceptual framework. Firstly, the generative is applied as a collection of individual

90


A MANIFOLD DIGITAL

techniques departing from the abstraction and transformation of processes originating in the natural sciences, which are transferred to practice by means of computational

and algorithmic principles. In this an interdisciplinary field between design-practice and computer science is formed. Secondly, beyond this concern with underlying techniques the generative suggests new organizational design principles where topological

schemas mobilize systems and their consolidation as form. This represents a conceptualization of design as concerned with process above form.

2. What I suggest to call the material and the generative tendency share a point of departure in a notion of a technologized nature. However the material tendency ap-

proaches nature differently. This assumes two general trends. One approaches nature concretely at the level of properties and behaviour of materials for building. This is

then utilized to drive and inform a design process. Another trend can be observed, where material properties are reproduced outside its initial context by way of bio-

mimicry. Both approaches suggest the use of digital media as a means to bring us closer to a material reality. Thus, the material digital tendency represents the move-

ment towards a concrete re-implementation of the digital in the real. I perceive the current material practices in architecture as directly necessitated by a need to rethink the relationship between structure, materials and performance in the wake of the formal novelty suggested by early digital architecture.

3. The tendency I attach to the notion of design realization relates to how digital mod-

elling, fabrication and planning today integrates to form intricate parametric assemblages representing a significant organisational shift in the nature of architectural rep-

resentation and their relation to their impending materialization. This shift has led to a new informational consistency in architectural design, where a complex information

loop collapsing the conventional stages of a design process allows a range of new pos-

sibilities for design – amongst others, the implementation of the generative and the material tendencies above within a total-process of architecture.

4. Lastly I observe a tendency within the many digital practices existing today sug-

gesting a condition of cultural reflexivity between the digital and the forces at work

in a cultural context. What is of importance here is how to arrive at new forms of expression that can effectuate a genuine cultural shift, or how cultural mechanisms

synthesize with technology to produce cultural change. It is a tendency towards relat-

91


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

ing to, and affecting, external meaning. This tendency is naturally related to a critical non-instrumental agency within contemporary digital design.

One commonality that certainly spans all of these practices is a profound interest in

advancing the discipline of architecture by looking beyond it. In order to develop ap-

propriate techniques a trans-disciplinary focus is always assumed linking architectural activities to fields and topics unfamiliar to a standard practice. Generative architecture utilizes computer scientific abstractions from biological morphogenetic processes and

other natural phenomena; material digital practices are, as already mentioned, seen to

reproduce material behaviour by means of biomimetic principles, but naturally also informed by material sciences in general; and recent trends in realization look amongst

others towards the aerospace and automobile industry for the development of complex information models. The Cultural reflexive trajectory can be seen as an attempt to at-

tach a broader significance and meaning precisely by means of the interdisciplinary within the interdisciplinary. A good example is Greg Lynn’s ‘Animate Form’ in which techniques borrowed from 3D animation software, combined with the mathematics

of topology and a Deleuzian philosophy challenges the concepts of stasis and permanence in architectural culture (Lynn 1999).

From the perspective of this field of tendencies, I will in the following direct attention

to the specific aspects within these tendencies that has significance for the project. My

method in doing so will be to exemplify concrete discourses followed by architects or theoreticians operating in relation to these. Consequently designating a more specific

context to which I seek to contribute. The examples are therefore not intended to describe tendencies exhaustively, but to define points of importance within each. In this I assume a certain perspective, in that a necessity exists to accentuate that the present

project is placed in proximity of the generative tendency sketched out above. Conse-

quently, this means that I observe and delimit the other tendencies from this position. The Generative as a Design Approach

The generative will be a recurring theme throughout the rest of the chapter, as it is also

within the thesis as a whole. As such I will add to the context of the generative while discussing the other tendencies as well. With this in mind I will in the following take

the opportunity to assume a broad perspective by which to observe how the generative

can be seen to define a certain procedural approach to design. That is, I choose to look

92


A MANIFOLD DIGITAL

beyond individual generative techniques towards the larger conceptual implications of

operating through such techniques. This directs attention towards a general challenge in generative design practice in understanding which procedures exist within a hypothetical “pre-material” state that mobilizes a movement into material reality. Benjamin Aranda and Chris Lash formulate it thus:

“If architecture is an extended process of formation, then before ideas coalesce

into a definitive form there must exist some undifferentiated state free of any organization. The moment any sort of development is imposed onto this formless

matter it begins to enter the realm of substance, organization, and material.” (Aranda Lasch 2006, p.8)

A generative approach can be observed to constructs a design process around the idea of departing from a topological schema defining the relations of a non-representational organization. This organization is a virtual construct allowing a variety of actualised

structures. As an example of this conceptual framing the architect Lars Spuybroek

defines two main phases of a generative design process, “a convergent phase of selection and a divergent phase of design.” (Spuybroek 2004, p.8) The first phase relates to move-

ments within a virtual continuum, as something diagrammatic, where information

is collected, selected and mapped. The end product is a virtual machine, instantiat-

ing movement towards quality, order and organization. The second phase relates to movements of actualization in the form of quantity, matter and structure, where the diagrammatic becomes formative.

The first phase is defined by research and analysis. Spuybroek holds that “[…] for self-generative design techniques we need empirical research of existing forms” (Spuybroek 2004, p.10). Although the generative require a topological outset, in order to generate a design it is necessary to topologies type. For example, if we are to generatively

produce a tower design we would need to know about the topology of towers in gen-

eral. From here on one can start applying a specific organization to the system that mobilises elements and relations within the system, as a specific ’machine’. That is, the

elements themselves start to assume an agency. From this agency the second phase commences by driving the system towards a concrete architectural morphology.

“We need to construct body plans out of this research through analysis, then these

machines must process information (or difference) through a mobilization of its

93


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

topologically connected components, which consolidate and take on a form, first as a design and then as a building“ (Spuybroek 2004, p.10)

In light of the discussion on optimization within the previous chapter it is noteworthy that the end result of Spuybroek’s self-generative procedural technique is stated to be a morphology of the provisional not the optimal.

“Often generative techniques are proposed under a sign of efficiency and optimization, but since the generative relies wholly on the topological and since the

topology is real and fully materialized, less determined (and more redundant) in-betweens are included.” (He continues…) “That means no geometry of

complexity, no morphology resulting from an epigenetic process can be fully Euclidean or elementary, because it is the relations that produce the elements, not

the other way around. The variability comes before the elementary. All shapes

generated through intensive processes are therefore transformative shapes and

have a transformative or, better, a transitive geometry.”( Spuybroek 2004, p.11).

Spuybroek’s approach exemplifies very accurately how the generative can be seen to

represent a concern within digital architecture that ventures beyond matters of indi-

vidual techniques. It is this view of the generative as an approach to the design processes as whole that I align the framework to.

For a better comprehension, it should however be noted that as a means of actualizing this approach the concepts of computation and algorithms are of importance in

understanding the generative tendency in digital architecture. These are the technological foundation that allows us to operate within the conceptual framework of the

generative; they are what technically mobilize elements and relations as generative machines, endowing them with agency. Computation relates to procedures by which

something is determined by mathematical or logical methods, it relates to “[…] the

exploration of indeterminate, vague, unclear and often ill-defined processes”(Terzidis 2006, p.xi). By means of computation we can formulate and utilize algorithms. Algorithms

are simply step-wise procedures or instructions defining how to approach a given task. However, through computation they allow “a complementary and dialectic relationship

between the human mind and the machine” (Terzidis 2006, p.37). This gives access to

degrees of complexity that would normally exceed our human cognitive capabilities. It

94


A MANIFOLD DIGITAL

[4]

4. Son-4 Son-O-House, Nox/Lars Spuybroek (2000-04) | 5.cv An analogue-computing model through which tectonics emerge from procedures of weaving and interlacing. For Son-O-House

95

[5]


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

is precisely this ability to operate beyond what human cognition allows that permits

certain digital design practices to challenge our notions of control and agency by way of the generative. Conceptually and technically the focus is here on process rather than representation, and formation rather than form. The design spaces of these practices

assumes the form of self-organizing systems often exhibiting emergent behaviour, orchestrated, rather than actually controlled, by a designer (Leach 2004). This orchestration depends on computation and algorithmic formulations. Material as Agency

Certain current digital practices can be observed to describe a gradient between the

generative and the material, where representations of concrete material components assume an agency driving a process of design. What can be suggested is that a material

tendency in architecture has developed into positions where the non-representational

character of the generalized generative approach above, through synthesis, has been

refined as a tool for simulating and mapping material performativity. This can be seen as a concrete manifestation of the technologized nature observed by Lash in the preceding chapter.

“Natural systems are now seen not so much as something from which to draw

formal order for design as offering a guide to how one can design with the perfor-

mative qualities of materials, entraining forces and material effects.” (Hensel, Hight, Menges 2009, p.10)

In this context architectural researchers Michael Hensel and Achim Menges propose an approach to producing architectural space that they term Morpho-Ecologies

(Hensel, Menges 2009). While ‘morpho’ might point towards a renewed focus on form as articulator of space, ecologies indicate that:

“the perpetual (and generative) interaction between objects and subjects and

their environment is a key concern. [Here…] material form and energetic flow folds objects, subjects and environment into an amalgam of effect-yielding inter-

action that is perceived as heterogeneous space” (Hensel, Menges 2009, p.196). While modernism in general favored a continuous homogeneous space delimitated by material boundaries, the Morpho-Ecologies approach is to instrumentalize environ-

96


A MANIFOLD DIGITAL

mental gradients between exterior and interior. In this way, the approach is to derive

material systems on the basis of their observed performative capacities. In order to use this approach to design heterogeneous space we are required to think of material systems as conditioners that generate and differentiate. These systems should not be

considered as static objects but rather as a process of material operations. The manipulation of such systems is necessarily connected to the modulation of environmental conditions. This observation renders possible an instrumentalisation of the interre-

lation between form, material, structure and a context specific interaction with the environment.

The approach repositions evaluation of material characteristics and behavior away from a rigid description towards one promoting responsiveness. This implies a certain

standpoint with regards to the nature of materials. Hensel notes that, while materialization throughout architectural history “was predominantly to do with reducing change and neutralizing its effect through some way of stabilization or compensation” today this

understanding is exchanged by a “notion of change of material properties and dimensions […] as a positive project” (Hensel Sunguroglu 2008, p.38). We should understand

material make-up, properties and behavior in relation to how we desire it to respond

when articulated at the meso-scale of the structure of a material system. In this lies an instrumental feedback between material system and environment.

For the Morpho-Ecologies approach this material understanding implies an interest

in the capacity of materials and material systems to self-organize in response to outside stimulus. Herein lies the reference to ecologies, understood as open self-maintaining

systems operating far from an equilibrium state, where form-giving influences derive from the system itself. Self-organization also underlies the method of form-finding we can observe in the material work of Antonio Gaudi, Frei Otto and Heinz Isler. Here

we are dealing with a design technique that utilizes the self-organization of material systems in such a way that the form is achieved by discovering the equilibrium state

of the system. Although Hensel/Menges acknowledges that this has mainly been used

for optimization purposes, they posit that: “Form-finding could be shifted from single-objective optimization to multiple-objective morphogenesis and be utilized differently in different regions or scales of magnitude of a material system.” (Hensel Menges 2009, p.208)

The suggested result of utilizing a Morpho-ecological approach is an architectural heterogeneous space that, firstly, allows the structuring and modulation of a field of

potentials that serves to provide a certain loss of formal and programmatic reference. Secondly, utilizes the reciprocity between material and environment to potentially

97


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

[6]

6. HygroScope: Meteorosensitive Morphology by Achim Menges and Steffen Reichert, 2012. A project exploring a novel mode of responsive architecture based on the combination of material inherent behaviour and computational morphogenesis

98


A MANIFOLD DIGITAL

project a transformation within a wider context (e.g. political or cultural). That is, an

emphasis on “process and the acceleration of the evolution of an architectural environment, in which the relation of form and space to program acknowledges the dynamic patterns of

human habitation” (Menges 2004, p.81). Environment is here understood as a dynamic gradient field of performative micro- and macro-milieus.

“Together these milieus produce an ecosystem, a dynamic relationship between environmental, topographical and structural intensities and human activi-

ties. Such an integral approach suggests that architectural design constitutes the modulation of micro-environmental conditions within an emergent macro-environmental system.” (Hensel Menges 2009, p.212)

This is understood by Hensel and Menges as a shift away from a unit-based design approach towards a condition-based. Recent technological advances allow us to further

develop such modes of spatial organization based on gradient conditions in interaction

with physical thresholds. We can now simulate the behavioral patterns of modulating, and being modulated by, an environment of a performative material system, even from early stages of the design process (I will return to this in the following section). Here we find a complexity that transcends a mere discussion of material system. The morphology occurring here is a complex system that is differentiated and performative.

The Morpho-Ecological approach is a move towards a more relaxed notion of control, as well as an intensification of performative potential and broadening of choice. It

supplies architecture with the possibility of propagation instead of spatiotemporal determinism, and a shift away from program as design defining towards design as program evolving. Program does here no longer designate a list of requirements but exist; “as a

post-design opportunity for human activities to be time-, space- and environment-specific on the basis of individual and collective preferences” (Hensel Menges 2009, p.212). Information Technology and the Total Processes of Architecture

When considering the total processes of architecture – from the conceptual to the build – current digital design technologies suggest a significant organisational shift in the nature of architectural representation.

“Where the first generation of digital design tools emulated the dimensions of

99


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

the drawing board, imitating the orthogonal logics of projective geometry and

presenting us with the known territory of section and plan, a new set of descriptions are fundamentally challenging the way that architectural representation is organized.” (Thomsen Tamke 2012, p.141)

This organisational change has resulted in a new informational consistency in architectural design, in which elements, knowledge and concerns traditionally found throughout the different stages of design development, now can be made to directly

feed into the early stages of a design. It is this consistency that I want to emphasise in situating the present project within a contemporary context of digital architecture. A complex information loop collapsing the conventional stages of a design process has

emerged, which presents us with a new open nature for the architectural model. I will

suggest three areas in which this fundamentally alters a concept of the model. Firstly, it assumes the form of a cybernetic type of system:

“How do we organize and articulate architecture in this ocean of information? The answer is obvious – by steering in relation to information, and navigating the bits“ (Klinger 2008, p.27).

Secondly, it emphasizes a notion of the model as a shared collaborative space of diverse expertise:

“[…] the digital model is interfaced with the world around it, receiving it in the

form of external data flows that describe the endless flux of its environmental, structural or material reality.” (Thomsen Tamke, 2012)

Thirdly, it allows the implementation of the material as an underlying presence informing the model signifying a new condition of a mediated digital craftsmanship:

“[…] the craftworkers have reappeared, only their focus has shifted from direct engagement with the material to creating information for materialization, digital fabrication, and assembly, in relation to material knowledge; encoding in-

formation is a form of craft that directs the craft of form.” (Klinger 2008, p.28) I will in in the following add some nuances and details to these three areas of interest.

100


A MANIFOLD DIGITAL

In discussing contemporary informational models architectural researcher professor

Mette Ramsgaard Thomsen and associate professor Martin Tamke suggest that these

represent a shift away from an image of the model as an autonomous object existing independently from an environment. Instead a perception of the architectural model

as an interfacing entity emerges, connected to a flow of environmental, structural and material information (Thomsen Tamke 2012). Here the model assumes a role where

it actively calculates what it embeds, in terms of contextual information, and in this

capacity takes on a hybrid appearance as tool and representation simultaneously – it

represents external meaning while calculating internally its own consequences with

respect to its context. This idea of the model as a design environment that calculates

what it embeds draws the nature of the model towards a generative and cybernetic understanding of design characterized by notions such as steering, agency and decentred control.

“To design becomes to instigate the agency of the model, the means by which the

flexibility of the design space and the mutable character of the designed can be

steered. Instead of understanding the designed as under a measure of control, what is suggested here is that the designed exists in continual dialogue with its

containing environment through which exchanges between design system and design environment are understood as a mutual negotiation. . Here, there is no primary cause or instigator. Instead, environment and system are conceived

as mutually dependent continually affecting each other through their recursive interchanges. � (Thomsen Tamke, p.153)

When contextualizing the thesis with respect to current trends in digital design, it is this notion of design as steering, decentred control and the model as agency that I wish to align the present work to.

In contrast to previous forms of representation a fundamental advantage of this active informational model of interface and flow is its ability to it establish a new shared

and collaborative design space. Within this space different types of expertise can become embedded within the same model. The diagram describing this condition is

not a solidifying entity in which a participating expertise is singled out as pivotal to the design process. Rather, the diagram represents continuously changing situations

where control is assigned to different degrees and where different types of expertise

101


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

[7]

[8]

[9]

7. - 9. Dermoid. A collaboration between CITA and the ‘Spatial Information Architecture Laboratory’, RMIT (2010). The project is realized through complex and interdisciplinary computational models allowing negotiation between production constraints, material behavior, gravitational forces, structural organisation and design intend.

102


A MANIFOLD DIGITAL

“can all claim the primacy of information as their driving force” (Klinger 2008, p.28). As-

sociate professor of architecture Kevin Klinger observes that this revolves around the

idea that a project is developed relative to a master model that iteratively is built from interactions within the collaborative. This type of model can be used in several interrelated ways; as a system of associations and constraints between related elements; for

testing the system by means of simulation and analysis; for describing the geometry to

directly fabricate the components of the final building; and to facilitate the eventual assembly process at the building site.

“The master model is the catalyst for enabling collaborative information exchange, which sets the stage for new structural conditions in the building industry.” (Klinger 2008, p.30)

In order to take full advantage of the active informational model Klinger holds that: “We must advocate for flexible structural conditions that enable fluid and direct information exchange in architecture […]. We must gravitate towards techno-

logically driven design through greater attention to research, experimentation, and production considerations. Additionally, we must encourage a total process

of design-through-production approach that engages all those involved in build-

ing design and production in a collaborative evolution of each project.” (Klinger 2008, p.35)

In extension to the material discussions in the section above, one significant aspect of the collaborative nature of contemporary information models is the potential role

played by the material. Here computation can be seen “[…] as a means of constructing a new relationship between representation and its immanent materialisation” (Thomsen Tamke 2012, p 141). Active information models can internalize an understanding

of their corresponding material systems. Here the material cannot be reduced to diagrammatic annotation, but exist as an underlying presence informing the model throughout its development.

“Digital making – the use of digital technologies in design and material production – is blurring the sharp discontinuities between conception and production

established in the twentieth century. New techniques based on close, cyclical cou-

103


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

pling of parametric design and digital fabrication are restructuring the relationships between design and production, enabling a closer interrogation of materials during the earliest stages of design.“ (Kolarevic 2008, p.120) Cultural Reflexivity

The examples described in the above, from tendencies within a contemporary digital architecture, have mainly been observant of changes and possibilities internal to ar-

chitectural practice. However, in extension to the discussion in the previous chapter

current digital architecture naturally also relate to questions of cultural discourse. My reading in this matter relates to the concept of reflexivity. Reflexivity of course im-

plies a two-sided relation. On the one hand, the question here is which mechanisms in digital practice might effectuate a change in surrounding culture and where these mechanisms reside relative to technology. On the other hand, I also find an emphasis

on how a surrounding culture drives and transform a digital practice. The concern here is where principles of change reside and is defined? Rather than attempting to answer these large questions I will in the following exemplify how they might be approached through the examples of others.

Change From an Outside

The architect and researcher Mark Goulthorpe expresses a reservation towards the promise of new possibilities for architecture allowed by digital systems. Although this promise seems genuine he observes that at the same time we encounter a difficulty

in establishing legitimate architectural principles aligned with such technological change. A potential inability to rethink the basic principles of operation within a new territory, “highlights the danger of constraining digital technologies to current expectations of

architectural form and to extant modes of praxis.[…] [Being a ‘digital’ architect] requires

that the new territory be thought as such, which suggests learning from those who already inhabit it effectively (programmers, mathematicians, etc.); these have long since recognized the

algorithmic, programmatic and parametric nature of such technology” (Goulthorpe 2003, p.180). Nevertheless this transdisciplinarity does not offer any qualitative notion in

itself. Goulthorpe notes that the result of solely limiting oneself to this path is that, within a digital practice, a desire for technology easily become the sole principle of le-

gitimacy. According to Goulthorpe this is in fact observable in a contemporary (2003) digital architectural scene that, in his words, has bifurcated between techno-rationalists on the one hand and techno-lunatics on the other. The former follow an ideal of a de-

104


A MANIFOLD DIGITAL

terministic mechanicism, the latter a vague formal speculation. While this formalist

branch does work towards a new practice in architecture, Goulthorpe stresses that a prescient imagination is required which is attuned to the spatial and social potentials

released by such systems. He: “[…] insists that it is in the yearning for hitherto “unin-

habited” cognitive space that a genuinely transgressive digital architecture might be born, imbued with formal, spatial or material quality.” (Goulthorpe 2003, p.180) What is re-

quired in order to accomplish this might according to Goulthorpe be an actual trauma. This concept of making the traumatic operative in an endeavor to effectuate a cultural

shift resonates somewhat with the project of another architectural practice; that of Francois Roche and partners under the name of R&Sie[n]. Roche’s work can often be

placed midway between architectural projects and art installation, often exhibiting what might be experienced as a provocative attitude. What R&Sie[n] do, is to make

phobic and other psychological concepts a mode of operation in architectural practice. The shapes of spiders, mosquitoes, excrement and hairy surfaces act as points of

departure for design (Corbellini 2009). What is more, Roche describes schizophrenia as a strategy of resistance, as something that allows you to speak from somewhere un-

predictable and to so with that appearance (Grobman Neuman 2012). Against what is observed as a rigid conformity in design practice justified as acts of performativity

accepting the established system (Roche 2012), R&Sie[n] “triggers a highly liberat-

ing action of disturbances, pointing to anomalies, and, above all, the construction of conditions that make them possible, as the necessary objective of architectural research” (Corbellini

2009, p.159). In order to understand this necessity Roche offers an example related to the use of algorithms in design:

“Any algorithm has a fundamentally linguistic dimension […but…] we do not

say ‘if, then, therefore’ all the time; we mostly settle for ‘maybe’ or for ‘perhaps.’

But it is difficult to integrate ‘maybe’ and ‘perhaps’ into computational language.” ( Grobman Neuman 2012, pp.102-103).

This kind of unpredictability comes from the kind of input driving the process. This necessitates a methodology that can produce an input that lies outside of the technological. It is to this end that R&Sie[n] develops their conceptual framing. Design

requires a kind of resistance against loosing autonomy to an instrumental reason and

therefore need to seek out concepts from outside such a reason. In this light, it is inter-

105


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

[10]

[11]

[12]

[13]

10. ‘OneMain Street’ by Mark Goulthorpe/dECOi Architects (2010) | 11. ‘Hypnochamber’, R&Sie(n) (2005) | 12. ‘I’ve Heard About’-project, R&Sie(n) (2005) | 13. ‘He Shot Me Down’-project, R&Sie(n) (2006-07)

106


A MANIFOLD DIGITAL

esting to note, that on their website ‘new-territories.com’ (www.new-territories.com, accessed June 4rd 2013) a statement informs that the site is, amongst others, dedicated

to ‘fiction as practice’.

Change From Within

In both the cases of R&Sie[n] and Goulthorpe mechanisms of change within a relation

of cultural reflexivity is to be found by introducing an otherness disturbing the tech-

nological. That is not to say that the technological is not an active part, but that change occurs in the synthesis between technological potential and a non-techological reflection. However, it is also possible to find examples where cultural change is claimed to be possible by the introduction of new technology alone. I believe Schumacher to

be an example hereof but another architect, Ali Rahim, whose work departs from the

same formal outset, presents a valuable argument exemplifying this concept. As he states in his book ‘Catalytic Formations’: “[…] I investigate [an] approach, […] that sees

digital technologies as platforms from which architects can develop new techniques, giving

rise to innovative works of architecture with significant cultural effects” (Rahim 2006, p.3). Feedback is a recurring theme for Rahim in this book, linked to the idea that innovation in architecture depends on architects and their work being more responsive to

their environments and users. This requires the incorporation of feedback from both a physical and a cultural context, as opposed to depending on conventional analytic

or internal processes. Rahim believes that the presence of new software and fabrica-

tion technology allows this feedback to occur, by which architecture can become more

dynamic, in the sense of being activated through interactions with users and contexts. According to Rahim this should make the work open-ended and as such supply it with a potential to realize unexpected effects.

Whereas Roche and Goulthorpe have an interest in the effect of culture on a tech-

nology-driven architecture, Rahim is mainly interested in how architects translate technological advances into innovative designs that produce lasting and significant

cultural effects. This ties to Rahim’s definition of technology, as a project of applying the technical, a concept aimed towards efficiency, to a wider cultural context (Rahim 2006, p.11). Thus, the focus shifts from questions of efficiency to a qualitative search for new patterns of behavior and levels of performance. The technical has no external

meaning without application in a context. The concept of technique is in this context; methods that allow individuals to use technologies in specific contexts, to accomplish

complex and difficult tasks. In order to participate in the process of cultural change,

107


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

the architect must seek to develop technologies and new techniques. Rahim proposes

four principles for evaluating the promise of new techniques: 1.) they should incorpo-

rate feedback from an outside, 2.) they should destabilize existing practices, 3.) they should be process driven (aim at continuous development), and 4.) they should add variety, which might come from other fields than that of architecture. Complying

with these principles, and resisting the use of technology merely to achieve efficiency, should according to Rahim result in ‘technological practices’ in architecture. This term is used to distinguish his work from ‘traditional’ forms of practice.

Rahim contrasts between analytical design-processes, possessing no time, and tempo-

ral techniques, which are fundamental to his concept of technological practice. These techniques synthesize the analytical methods used by conventional practices with bottom-up approaches. They are non-linear and combine existing entities in such a way that they produce new and emergent organizations that are more than the sum of their

parts. They combine numerical and virtual components. Thus, temporal techniques include both factors that are controlled by the designer numerically and ones that

emerge in unforeseeable ways within the system. Consequently, temporal techniques are characterized by being generative and transformative.

Technological practices are especially interested in producing affective formations. This has to do with their definition as responsiveness to users and context. Rahim exem-

plifies this with the image of the street curb, which is an effect caused by the intention to separate pedestrians from automobile traffic; but when someone uses it as a bench

this is an expression of the affect excised by the curb. For architecture to generate cultural change it must be engaged in a relationship of mutual feedback with its users and

contexts, and in this possess the capacity to affect and be affected. Affects inform the

process of formation and vice versa. Techniques, affects, and form respond to and alter one another, constituting a feedback loop. According to Rahim affects are embedded in the formal properties produced by transformations generated as part of temporal techniques. These transformations generate affordances; a concept which designates

specific properties of formation that indicate how one can interface with the formation. Thus, affordances activate the affects. Because affordances are activated by us-

ers, this generates responses that according to Rahim may feed forward to instigate broader cultural effects.

108


A MANIFOLD DIGITAL

[14]

[15]

14. & 15. Ali Rahim, Reebok Flagship Store, Shanghai, China (2004 - )

109


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

Aligning the Framework

In establishing a context for the research within a digitally designed architecture I position the work in a specific historical lineage of architectural discourse. This lin-

eage departs from a break with modernist as well as post-modernist discourse in the form of a shift towards a topological mode of thinking. I do not observe this lineage to have fashioned the field of digital design in the form of a coherent movement

within architectural discourse. Instead I observe it to have proliferated into a range of tendencies explored to different degrees within individual practices. In position-

ing the thesis relative to these tendencies I am able to account for how its proposed framework situates itself within a total context of digital design. Here the framework observes the generative as its approach to design; material as a specific agency within

this approach – assisting in coupling architecture to its environment; the design space as a non-linear interdisciplinary system of internal exchange; and the design itself as a mechanism of cultural reflexivity.

The Generative as Approach: The framework is naturally positioned in close proxim-

ity to a generative perspective for design. What is important to state in this matter, is that the framework is less concerned with individual generative techniques, as with

the generative as a specific procedural approach to design, as exemplified above. Tech-

niques are important but they are a priori conceptualised by the generative in this role. Here the generative supplies an inter-disciplinary morphogenetic perspective on design directed by a non-representational principle of organization and form-develop-

ment. Assuming this view, beyond individual techniques, allows the generative to form an understanding of the design process, which can also be made to have significance for other active tendencies in the field.

The Material as Agency and Interface: Architecture is by definition a material prac-

tice. Consequently a framework for design should acknowledge this characteristic. A current synthesis can be observed between an architectural concern with materials

and a condition where architectural design begins to constitute the modulation of local conditions within a global environment. Thus, the presence of material concerns

allows a framework for design a point of interface between the generative and a concrete physical, cultural and societal context. What is specific to a generative approach

to architectural materiality is how the material is seen to assume agency. This agency

110


A MANIFOLD DIGITAL

can make the generative context-specific. Material performance and responsiveness

become part of the organizational principles of design which constitute a gradient field of exchange between interior and exterior. This interface of material agency, ar-

ticulating the architectural structure within an environment, constitutes a decentered mode of control, which evolves rather than prescribes a design.

Representation and Informational Consistency: The contemporary information

model is of interest because it ties the generative and material approaches above to current trends in the complex relation between design, fabrication and planning. Although these models are not contingent on a generative approach as such, they can

be seen to exhibit generative characteristics. The informational consistency observed in these models endows them with the potential for a generative complexity, where

design is to activate the agency of the model. What is of interest is the potential for the generative to participate, and be driven by, the interdisciplinary knowledge inter-

nalized within a complex information model. This notion of design as a system that extends beyond disciplinary division, in time as well as space, relates directly to the topic of cybernetics discussed in the following chapter.

A Coupling to Culture: In order to qualify and legitimize architectural principles a framework for design is required to observe how it interfaces with the culture to which

it seeks to contribute. In this regard I subscribe to the position that the facilitating mechanisms driving a system for design must lie outside the technological. This is consistent with rejecting to operate relative to an instrumental reason as argued in the

previous chapter. However, I also acknowledge the claim of Rahim, that a technologi-

cal practice is not neutral in its relation to its cultural context. Thus, the design space assumes the form of a mechanism for cultural reflexivity. It assumes an overall form

reminiscent of an algorithm: its unfolding is dependent on its environment, but represents its own internal rules.

111


5 | A CYBERNETIC APPROACH


A CYBERNETIC APPROACH

While the previous chapter explored a broad range of characteristics related to a contemporary digital discourse, I will in the following start a movement towards more

specific aspects and understandings of the particular framework for design suggested by the thesis. In this regard I will begin this chapter by sharing a number of observations made while working on the Lamella Flock project. I will use these to begin a discussion related to design methodology, in light of the specific technological and

technical context in which the project is engaged. The Lamella Flock project will be

discussed in greater detail following this chapter; however some introductory remarks, of a general nature, are required in order to understand the specificities of relevance here.

By means of introduction, Lamella Flock investigated the translation of a set of traditional wood construction principles – the Zollinger construction – into a system for

a non-standard practice. The original construction principle behind this system was characterised by a logic of circular dependencies, which required a non-linear mode

of making the system operational in the design of free-form constructions. What is of interest in the present context is the type of digital design space that developed

as a solution to this challenge. This design space circled around the idea of implementing systems having their own agency, as a technique for developing architectural

form. More specifically, the prototype structural system was re-conceptualized within a computational model, as a collection of time-based entities with a specific autonomous behaviour. This behaviour allowed elements to reconfigure themselves in a negotiation with other elements in accordance to the geometric principles of the original system, as well as in accordance to restrictions related to their eventual production as

physical structural elements. In working with the development of this agent-based

model I observed a number of significant aspects tied to this type of design space, which I find valuable in a continuous discussion of the role of the designer and the nature of design itself. That is, they suggest something about design methodology and practice, which I attach to the framework for design described through the thesis.

In designing the Lamella Flock system, as an agent-based solution, I worked primarily

through the use of a programming language, i.e. not through representational techniques. Namely, I developed the algorithms by which something representational took

form, not the form itself. In doing so I discovered a certain discrepancy between coding as planning and its result in the shape of what I observed on the screen. I found it

113


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

[1]

1. Lamella Flock (2010), generative design system. A series of screenshots showing the process of formation by the agents

114


A CYBERNETIC APPROACH

hard to analytically predict the exact behaviour and interaction of the agents, because

of their complex non-linear interactions. Still, this did not prevent the design of the system to progress. Instead I would iterate between altering the code and observe the time-based result, as it played out on the screen. Thus, a process developed in which I

could be said to observe the performance of elements within the experiment, and this

observation would feed directly back into the further development of it. The diagram by which I understood the use of computers changed. From a somewhat naïve ideal

that we can have a full representational control of the workings of a complex system,

where the computer act as interface between a human input and a desired output, the system by which I worked became one of learning and observing in a non-linear fashion.

This was what I experienced in the developmental phase of the project, but the final

design space of the Lamella Flock project suggested that many of the same character-

istics remained in the process of designing with the system as well. For anyone using the system – which allowed interference and interaction with the individual agents – a

process of design was a process of learning and observing. One would initialize the agents in a certain way and then manipulate them locally, while the global behaviour resulting from the interacting agents produced a complex negotiation leading to a coherent structural solution. To design successfully with this system required a continu-

ous process of gathering knowledge of the local and global behaviour of the system through observation and interaction. In this, design solutions were arrived at, neither

through the agency of the model, nor that of the designer, but through a decentred

shared agency. Again this was not an analytical process, but rather a performative, where knowledge and information was partly externalized from the designer.

These experiences were gathered within the scope of an isolated project, but I will

suggest that two observations appear in contemplating it, which I wish to extend to the description of the framework for design proposed by the thesis. Firstly, conceptualizing the digital design process as planning gave way to a process of indeterminacy and performance. Secondly, in this, the role of the designer turned from being that of a predicting and imposing agent, to an observing and learning mediator.

In an effort to find a way of conceptualising across these two methodological observations – in order to point towards something broadly applicable to design – I will in the

following make suggestions as to how the field of cybernetics supplies an appropriate

115


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

conceptual model for doing so. What we are supplied with here is a coherent frame-

work into which to place concepts such as observation, agency, control and reflexivity. In addition I will also show how a cybernetic model points towards aspects within

preceding chapters. Firstly, by way of cybernetics I can expand upon the discussion of the application of technology and science begun in chapter three. Here the application

of technology was made to assume a role beyond that of an applied science – as some-

thing by which we can pose question beginning with ‘how’ as opposed to the ‘what’ of a conventional science. As will be shown cybernetics complies with the former

and opposes a conventional modern science. Secondly, I will show how cybernetics

contributes to aspects of the digital architecture I highlighted in the previous chapter – the generative as an approach, the material as agency, the design space as a complex

informational system and the reflexive relation between architectural design and the culture in which it resides.

With regards to the structure of the chapter, I divide the following up relative to two

main discussions. To begin, I investigate the question of compatibility between design

methodology, cybernetics and the project of the thesis. Here I will briefly define my

understanding of cybernetics, followed by discussing the idea of conceptualizing design as a cybernetic system. After this general development of the framework I will highlight three concepts from within the field of cybernetics, which can be used to expand upon the experiences described above as well as related design spaces.

Compatibility: Design as Cybernetic System

The overall aim of the following is to discuss cybernetics in relation to design meth-

odology. I would like to demonstrate how cybernetics can be shown to relate and

contribute to the activity of design – partly in a general sense, but most importantly, specifically in relation to the framework proposed by the thesis. In this endeavour what I would like to point towards, as the particular contribution offered by cybernet-

ics, is how it can be utilized to define the role of the designer. However, before doing so

I need to specify certain concerns with regards to the concept of design methodology, and to situate my understanding of the field of cybernetics.

Design Methods

The development of design methodology is a problematic topic, especially when un-

116


A CYBERNETIC APPROACH

derstood as a prescriptive and generalising project. What I would like to address in relation to such projects, and in light of the topic of cybernetics here, is a pursuit of

attempting to improve design activities through scientific study. Historically this can

be exemplified through the post war ‘Design Methods’-movement, where methodology was seen to be something that prescribes to a design process a stringent procedure

which systematically collects information, establishes objectives, and computes the de-

sign solution – a procedure that follows the principles of logical deduction and math-

ematical optimization techniques (Gedenryd 1998). Instances hereof can be found in the form of Christopher Alexander’s ’Notes on The Synthesis of Form’ (Alexander 1964) and John Chris Jones’s ‘Design Methods’ ( Jones 1970). In a comprehensive study

Gedenryd shows that design methods all follow a similar pattern (Gedenryd 1998). These methods emphasize design as planning. The belief is: that by means of a process

of analysis we can specify a path backwards from a known goal, designating the steps followed by a process of synthesis – moving in the opposite direction – to produce that

goal. Thus, the design process becomes reversible, linear and symmetrical in relation to its product. Here action becomes an extension of rational thinking, and the plan

the construct that connects these two. The main point of failure lies in thinking that analysis and synthesis can exist as pure phases within the design process. Gedenryd suggests that in fact they are two aspects of the same activity, and this activity is one

that proceeds iteratively in cycles between each (Gedenryd 1998, p.55). It is only as a means of communication, after the event, that a design process can be shown to follow distinct phases of synthesis and analysis, leading to a solution to a problem. But even

this notion of a ‘problem’ is misleading. In design problems are just as much produced, by the designer, as they are given. Inventing problems is also an iterative process between the designer and the product as it materializes.

Now, what is actually at stake here is that design becomes methodologically connected

to a conventional modern scientific practice. Thus, the prototypical design methodologies of Alexander and Jones are closely reminiscent of how mathematics and formal logic operate. The intention of borrowing from science is of course to supply the practice of design with a robust foundation for making good design decisions, within an

efficient framework for the design process. Even if one can sympathize with such a project, incompatibility between methodological models of design and science makes this fail. But what if other models, with better compatibility exist? Before considering this question, a few remarks need to be made:

117


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

First of all, I believe that the idea of a universally applying prescriptive design methodology is in conflict with the ideological view arrived at in chapter three. Accordingly, a

description of any kind of methodological model for design has to have its procedural basis within the locally determined, autonomous and individual.

Second, when a certain human agency for design seems to exist well on its own, why then consider synthesizing with methodological models outside of design at all? I

believe that the answer is twofold. For one, design is dynamically coupled to the differ-

ent contexts to which is seeks to contribute. Thus, there are no meaningful boundaries defining a context-independent design process, and therefore a continuous reinven-

tion of methodological models makes sense. Also, in the concrete circumstances of this thesis I am combining elements that are already introduced from a context outside of

design: e.g. mathematics, computer science, complexity theory, biology etc. In order to implement and combine such elements, methodology must be considered. One cannot relate these to design ‘as done’ because they are at most peripheral to this context.

(In)compatibility

As just described a modern science, on the one hand, and design on the other can be argued to be methodologically incompatible. However, earlier I introduced the idea of two kinds of science. These were described by Deleuze and Guattari as royal

and nomad sciences. While the thesis itself consolidates itself between these two, I

believe that design in general, and therefore design within the framework, tend towards the nomad sciences. When the above described design methodologies fail, it

can be hypothesized to occur because they attempt to project a royal science onto a nomad, rather than acknowledging a situation of exchange between these modes of

operation. Within a contemporary digital design practice, we are certainly in a continuous process of appropriating technologies from the hard sciences, even when our aims are somewhere else. But instead of insisting on borrowing models from these

scientific domains, in order to further a methodological framework, I will posit that when design needs models for strengthening its methodological foundation one way to proceed is by looking towards other nomad sciences.

This finally leads us to the topic of cybernetics. Sociologist, philosopher and historian of science Andrew Pickering proposes that the field of cybernetics can be theorized as a nomad, or minor, science and for that matter also an anti-disciplinary and nonmodern.

118


A CYBERNETIC APPROACH

“We might think of the cyberneticians as nomads, and of cybernetics as a nomad

science, perpetually wandering and never finding a stable home. […] The no-

mad sciences, on Deleuze and Guattari’s reading, are a different kind of science […], one which wanders in from the steppe to undermine stability.” (Pickering 2010)

Importantly, this is not to say that the philosophy of Deleuze and Guattari subscribes to the principles of cybernetics, only that cybernetics complies with the definition of a nomad science.

An Understanding of Cybernetics

As a nomad science, cybernetics can historically be described as characterized by an

inability to be contained within disciplinary, ideological and institutional frameworks. Therefore it is also a slightly fleeting topic to approach. However, combined with

the notion of being comprehended as nomad, I will in the following account for the

general lines along which cybernetics is understood within the thesis. Firstly, I will begin by pointing towards the usefulness of cybernetics, as a field that illuminates the

concepts of adaption and control by way of abstraction. Secondly, I will point towards

cybernetics as a practice of re-implementing abstracted mechanisms as tools, models and devices within new contexts. Lastly, I will address the significant cybernetic principle of an observer, which is of particular value to an understanding of the role of the designer in light of the experiences described in the opening of this chapter.

Mechanisms of Adaption and Enabling Control

In essence, cybernetics is a conceptual framework concerned with those properties of

a system that are independent of their material or component parts. In this regard,

what is of interest to me is how this potentially allows adaptive mechanisms to ‘flow’ between contexts. By looking for isomorphisms between systems, and through ab-

straction from physical aspects, cybernetics investigate and describe physically very different systems; economies, brains, eco-systems and electronic circuits, amongst others. The general concern of the field is how such systems use information and control-

actions to steer towards and maintain goals, while adapting to a dynamic environment (Heylighen and Joslyn 2001).

This concept of steering, central to all cybernetics, should be observed from the per-

119


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

spective that the focus is on systems where information, control and communication are intended to operate jointly together. Since such systems are found in, as well as

outside, nature, cybernetics represents an unprecedented synthesis of the organic and

the mechanical (Hayles 1999). This points towards an important, though somewhat

problematic, aspect of cybernetics. The initial purpose of the field was to formulate central concepts that would merge into a theory of communication and control ap-

plying equally to animals, humans and machines (Wiener 1961). In this we might be

led to believe that cybernetics is a project of instrumentalising organic beings, as if they were machines. I believe that this is mistaken. What is important to note, is that

the concept of control, found at the heart of all cybernetics, is to be understood in the

sense of enabling control, not the restrictive control imposed on something or someone (Glanville 2007). In this the purpose of cybernetics is to allow better performance in

a broad range of situations, and to fashion machine more in the image of man, than making man function like a machine (Hayles 1999).

A Practice of Adaptive Modelling

Andrew Pickering proposes an entry into cybernetics connected to a notion of how the adaptive mechanism above finds their way back into a material reality. It takes its

point of departure in conceptualizing cybernetics as a project that originally reflected

on what a brain is. Cybernetics opposes a traditional view of the brain as a container of representations, an organ of knowledge or a thinking machine. Instead cybernetics

suggests that the brain is an acting machine, an embodied organ, intrinsically tied to

bodily performances. The purpose of these performances is to make us adaptive within changing environments. In this Pickering defines cybernetics as a science of the adaptive brain (Pickering 2010). However, Pickering quickly emphasises that cybernetics as a developing field cannot be contained within this definition – partly because of its interdisciplinary nature, but also as a consequence of the fact that the ideas proposed

by cybernetics requires experimentation. Historically this can be seen to have led to a

practice of building actual devices in order to understand the adaptive processes of interest. Although these devices originally tried to explicate the functioning of the brain

as a complex adaptive system, they can also be seen as things in themselves; functional

as technologies and tools beyond their capacity as models, which is a point of interest for the thesis. In this, cybernetics can in fact be understood as more anti-disciplinary

than interdisciplinary, and as such quickly expands into robotics, engineering, general

120


A CYBERNETIC APPROACH

systems theory, biological computing, management, politics, entertainment, the arts, theatre, music, education and architecture (Pickering 2010). The agent-based modelling of the Lamella Flock project, as described above, links directly to the adaptive devices of cybernetics, which I will return to below, and the agency of the designer to the above perception of the brain as an acting machine.

Second Order Systems and the Importance of the Observer

In relation to a discussion of the role of the designer, one additional topic needs to be

added to this brief introduction to my understanding of cybernetics – the concept of the observer. As already explained cybernetics was from its beginning interested in the

similarities between autonomous living systems and machines. Departing from new

control and computer technologies, cybernetics started out developing this interest through an engineering approach, where the person designing a system determines what it will do. Later on, a desire to distinguish cybernetics from these more mechanis-

tic approaches occurred. This took place by emphasizing autonomy, self-organization, cognition, and the role of the observer in modelling a system. This change within the

field is usually referred to as second order cybernetics. The change occurred as a paradigm shift in which cybernetics began to doubt its own nature as a traditional scientific

discipline. What transpired was that cyberneticians began to consider the idea of the observer within the system, as opposed to the observer of the system. This presents

an incompatibility with the detached observer of traditional scientific methodology. ‘Second order’ refers to the fact that the observer of a cybernetic system is here itself such a system; wherefore a new second order system exists between the two. Whereas

first order cybernetics placed the goal outside of the system, second order cybernetics positions the goal within, because this is where we find the observer. This also means

that a second order system become incomprehensible from an outside position, while the first order system can become somehow objective in a more conventional scientific sense (Heylighen and Joslyn 2001).

The role of the observing system is also one of control. This concept lies at the heart of

cybernetics, but as noted above in the sense of enabling control, not restrictive control. What is of concern with regards to cybernetic systems is what constitutes control in a system that enables rather than restricts? William Ross Ashby, one of founding figures of cybernetics, defines the answer to be, that for control within a system not

to be restrictive, the system for control must have as many states as the system being

121


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

controlled (Ashby 1956). But it must be noted that within a (second order) cybernetic system the controller and what is controlled are related by circular causality, which

makes it difficult to distinguish one from the other. So in fact control can be described

as shared. This radically separates cybernetics from science where circularity is a concept targeted for elimination. The controller within a cybernetic system is an active and involved observer.

Design and Cybernetics

I will now return to the main aim for this part of the chapter – the possible rela-

tion/merge between design and cybernetics. In this endeavour I will investigate an

argument in this matter made by architectural researcher and cybernetician Ranulph Glanville (Glanville 2007). This argument should be seen as an extension to the above

remarks on design methodology and its use of scientific ideals for architectural practice.

Historical Parallels between Design and Cybernetics

There are certain similarities between the development of architectural design prac-

tice and the field of cybernetics, amongst others in their relationship to science. In an attempt to show that second order cybernetics can act as a theory for design and design as a model for cybernetics in practice, Glanville insists that design can never be subordinated under scientific standards (Glanville 2007). Design is not a science, but

a way of acting as well as thinking. But, as already touched upon, certain tendencies in design theory can historically be demonstrated to have been searching for ways to turn

the activity of design into a scientific pursuit, complying with a strict and replicable

methodology. Cybernetics was at an early stage seen as a means through which to do so. Alexander’s ‘Notes on the Synthesis of Form’ was inspired by Ashby’s first book

‘Design for a Brain’ (Ashby 1960). This was a plausible idea since first order cybernetics assumed a much more traditional scientific nature than the field would eventually

have. As noted above, systems complying with this type of cybernetics project goals

from an outside, necessarily resulting in a mechanistic and deterministic thinking. This is evidently compatible with an ideal of making the system of the design process operate like a machine with input and output.

However, the development of both cybernetics and design theory documents a tendency towards failure in an endeavour to comply with strict scientific standards. In

122


A CYBERNETIC APPROACH

relation to this failure in the design discipline I have already mentioned the work of

Alexander and Jones. Alexander himself acknowledges the failure (Alexander 1971). This realization occurred around the same time as cybernetics also began to question its own nature as a traditional scientific endeavour; when it began its transformation into its second order equivalent and assuming its current form as a way of thinking linked to (inter-)action and observation.

As the field of design abandoned the ideal of the scientific universal design methodology, according to Glanville it as well turned towards something similar to the idea of an embedded observer (Glanville, p.1177). In this particular case the presence of the individual designer within the design process. The result hereof was a tendency

towards expressing theories as (personal) styles, simplifying the work-context in ways

where many design decisions came to exist a priori. Glanville notes that both cybernetics and design thus accepted the inescapable presence of the observer, but that

design somehow failed to notice the change in cybernetics that made it specifically concerned with understanding systems in which the outcome is unpredictable and

individual. In this role it can be suggested that cybernetics is still of relevance to the design profession.

Design as Cybernetic Conversation

The design process is necessarily based around the actions of a designer, so it would

be illogical to talk of this process as if the designer is not present in it. This acknowl-

edgement has the potential to direct us towards a cybernetic mode of thinking about design. The playful ‘conversation’ taking place between the designer and the object

being designed, interactively altering the outcome as requirements are gradually assimilated into the design, indicates a cybernetic process at work. For example sketch-

ing can be described by means of a primary second order cybernetic system, as a circle of conversation. This concept originates from Gordon Pask (Pask 1975). Conversation

involves listening and speaking in an essentially circular form. Pask held that conversation is the basic form of genuine interaction, herein lays its importance as a model

for design. We can conceptualize design as a kind of (cybernetic) conversation. This of course means that we should understand listening and speaking, as analogous with viewing and drawing. Within conversation a difference occurs between listener and speaker, and it is within this difference that novelty can be produced. Glanville states

that it is through this difference that the variety of the ‘repertoire’ of the designer can

123


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

be increased (Glanville 2007, p.1190). The link to conversation theory can also be

utilized in order to demonstrate the importance of failure. This occurs when we fail to complete the conversation circle. When this is observed we gain an insight that

allows us to restart the process, and in this way failure is not understood negatively, but productively.

When failure nonetheless does not occur, how do we know when to stop the pro-

cess, when have we arrived? The area in which a designer operates is ill-defined and

therefore an endpoint is only recognizable through the concept of adequacy. Glanville suggests that here we can utilize another type of cybernetic system as a model: Von

Foerster’s concept of Eigen forms (Von Foerster 1977). The interest of Von Foerster

is with systems following processes that, disregarding starting point, always end up and continue to be at the same place. In a more technical language, we are to think of

recursive functions that stabilize on particular values. They model the coming into being and a persisting to be. These processes reach a point where they start to reproduce

themselves iteration after iteration. Similarly the design process eventually reaches a

point where new iterations reproduce the form of the previous iteration. This is the point where we arrive at a solution. This is not necessarily the best solution, since this

concept makes little sense within design, but an adequate solution. In the end the question is to “create the conditions in which the design outcome can come into being and continue to generate itself ” (Glanville 2007 p.1192), i.e. to persist.

Complexity and Solution

Glanville define design as an effective approach to complexity, because it defines problems through its solutions rather than allowing problems to lead to a solution. Op-

posite science it is inductive rather than deductive. Its effectiveness lies in the fact that problems addressed by design are often so complex that their interrelationships quickly lead to a problem space with a complexity beyond computability. Meeting

these problems ‘the other way around’, i.e. after the event, allows the task to become manageable, but this also means that solutions can never be validated against a measure of success related to a best solution, but rather one that is good enough. The latter

would anyway require that we positioned design within the realm of the predictable, and this is incompatible with the aim of design as creation of something new, which

is by definition outside of prediction. Previously I mentioned how Ashby states that

non-restrictive control requires that the variety of states within the controlling system

124


A CYBERNETIC APPROACH

must never exceed the variety in the system to be controlled. But when we operate

outside prediction at a complexity where variety is beyond computability we must necessarily lose the ability to control. In this way design problems are in fact unman-

ageable. But this is in fact not necessarily a problem, since it opens up the possibility of

discovery within this unmanageable variety, because we have to abandon a restrictive control. Operating relative to unmanageability is a way of increasing our creativity. It gives access to ideas, otherwise not discovered. Glanville holds that much of his argu-

ment is dependent on a decision not to control. A conversation is not a conversation, if controlled by one of the participants.

Although design operates relative to complexity, the goal of the outcome of a design process is not complexity but simplicity. What is needed in this regard is a process by which complex requirements can be brought together within one unified form. For Glanville this is the goal of defining design as a second order cybernetic system: “Rather than try to specify every requirement and every relationship between these requirements, and then find an optimal solution, design starts more-or-less

“aimlessly” and gradually constructs an “evolving” form that not only changes, but in doing so accommodates the required functions also, often in a novel and surprising manner, where normal relations between functions are enriched or

even replaced by new ones that are unexpected, different, and often very good! The accommodation of further requirements within the form, the assimilation

of requirements by substantial shifts in that form (including rejection and re-

starting), and the conversational manner in which this is done (itself leading to developments in the form) all help designers to simplify the complicated (com-

plex) in finding their final form through a recursive, circular action process.” (Glanville 2007, p.1196)

Glanville defines design as the quintessential constructive activity. Designers construct

new realities. Epistemologically: “the act of design is constructivist and the analysis is sec-

ond order cybernetic. In fact, design is perhaps the most universal and widespread of all sec-

ond order cybernetic activities“(Glanville 2007, p.1199). The difference between science

and cybernetics, in analysing a system, is that science gives us knowledge of the system, whereas cybernetics, and therefore design, produces knowledge for acting within the system. The former represents a passive, neutral and detached knowledge, the latter, an

125


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

involved and active knowledge. In this way cybernetics coincides with Lash’s view on technology in chapter three, as something by which we might pose questions beginning with a ‘how’.

Synthesising with Cybernetics I

The above shows that cybernetics can be utilized to define the role of the designer. It supplies a conceptual model by which we can choose to understand a system for

design. Although this relates to an understanding of design in general, I believe that key aspects of cybernetics are particularly applicable for conceptualizing across the experiences of the Lamella Flock project. Thus, offering the outline of a conceptual

framing of the specific design spaces relevant to the thesis, including those described in the previous chapter. Lamella Flock should here be understood as the placeholder of a more general practice of architecture. In closing this part of the chapter I will in

short form restate the key findings of the previous sections by directly linking them to the design space experienced within the Lamella Flock project.

A clear in-compatibility can be seen between the historical design methods movement

and my experience of an agent-based design system. Namely, that design as planning give way to design as a reflexive activity. This failure occurs by attaching the ideals of a

modern science to the practice of design. Instead I suggest that compatibility between

science and design can be achieved by turning from a modern (major) science to a nomad science. Cybernetics represents such a science.

Cybernetics can be said to represent the practice of a double movement by which

it appropriates mechanisms for steering by way of abstraction, while simultaneously, re-introducing these mechanisms as concrete devices, tools and models within new contexts. As a practice the agent-based modelling of the Lamella Flock project parallels this double movement. The agents are based on adaptive mechanisms abstracted

from nature, which are reconfigured in order to produce a solution to an architectural

challenge. Accordingly, at the level of technique the project points towards a cybernetic practice.

This assumption is further emphasized by how my role as a designer, within the de-

sign process, can be described. In direct parallel to the observer within a second order cybernetic system, I am an observer within the system of the design space; both in

the process of designing the system, and designing with it. From within this system

126


A CYBERNETIC APPROACH

I attempt to control a process. However, this is done by mode of an enabling form of control, whereby I am able to steer within the system; all the while the system itself assumes an agency by which it exercises a degree of self-control. Thus, occurs something

analogue to the circle of conversation of Pask, now described as a process of iteratively coding, manipulating and viewing. Similar to the observation of Glanville the occur-

rence of failure is important to this process, where agents will often ‘misbehave’ in ways which makes a design solution impossible, but simultaneously become the mechanism driving the design forwards as a process of learning. Designing the system, as well as designing with the system, also exhibited a characteristic similar to the Eigen forms of

Von Foerster, unfolding interactively but eventually stabilizing on particular ‘values’ – that is, as a stable solution. Importantly, in the case of the Lamella Flock project, this

solution was a solution to an unmanageable and underspecified problem, to which its cybernetic approach allowed the means to arrive at.

As I stated in the beginning of this section, investigating the Lamella Flock project

as a cybernetic system has the purpose of offering the outline of a conceptual framing for the specific design spaces relevant to the thesis. Thus, I suggest understanding the

described conceptualization of the role of the designer, as a component to which the

framework for design – established throughout these pages – aligns. Consequently, the design approach of the framework is defined as cybernetic.

Contributions: A condition, a space and a mechanism

Having argued that design can be conceptualized as a cybernetic activity and the de-

sign space as a cybernetic system, I will in the remainder of this chapter present a set

of conceptual components from the field of cybernetics that I find relevant to intro-

duce within a framework for design with the characteristics of relevance to the thesis. Here I identify three topics of interest and in direct extension of the first part of the chapter. Firstly, a condition of decentred agency attached to a certain productive space

with consequences for the question of representation; secondly, a non-representational approach to navigating an environment; and lastly, the operative mechanism of reflexivity.

A Decentred Agency within an Aesthetically Potent Environment

Cybernetics implies new models of subjectivity, thus, implicitly also new models of subjectivity for the designer, as accounted above. Hayles describes how the boundaries

127


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

of the autonomous subject are contested in a cybernetic perspective (Hayles 1999). Information does not only flow within the subject but between the subject and the environment, and so emerges what Hayles defines as the posthuman. The concept of posthuman is meant to be suggestive rather than prescriptive. In its extreme it

privileges informational patterns over material instantiation; it observes consciousness as an epiphenomenon, not the total picture; it observes the body as the original

prosthesis that can be replaced or extended to other prostheses; and it configures human beings so that they can be seamlessly articulated with intelligent machines. In a posthuman perspective there are no absolute demarcations.

“The posthuman subject is an amalgam, a collection of heterogeneous compo-

nents, a material-informational entity whose boundaries undergo continuous construction and reconstruction. “(Hayles 1999, p.3)

I believe that it is not within the scope of this thesis to enter into a discussion of the

posthuman aspect of cybernetics per se. This topic would require a thorough ethical investigation as well as the need to venture into speculative technological discussions

concerning human/machine synthesis1. I will instead suggest the more humble con-

cept of a condition of decentred agency, as an insight offered by cybernetics, that we

might consider being of relevance to a design context like the present. Here I would like to introduce the work of the cybernetician Gordon Pask as a key insight. Concluding his essay The Architectural Relevance of Cybernetics, Pask describes the role of the designer within such a condition:

“[…] the design goal is nearly always underspecified and the ‘controller’ is no

longer the authoritarian apparatus which this purely technical name commonly

brings to mind. In contrast the controller is an odd mixture of catalyst, crutch, memory and arbiter. These, I believe, are the dispositions a designer should bring

to bear upon his work (when he professionally plays the part of a controller) and these are the qualities he should embed in the systems (control systems) which he designs.” (Pask 1969, p.496)

According to Pickering, the defining topic of all Pask’s work was the concept of an aes1. For a discussion on this see for example Hayles 1999

128


A CYBERNETIC APPROACH

thetically potent environment (Pickering 2010, p323). This concept describes a system

that ‘man’ – seen as an adaptive being – performatively interacts with in an open-ended

exchange to produce novelty. Pask defined a set of conditions required for a system to

be an aesthetically potent environment: It should offer a sufficient degree of variety, so as to neither be too undifferentiated nor unintelligible; it should contain forms that

a man can learn to interpret at various levels of abstraction; it should provide cues in order to guide a learning process; and in addition, it should be made to respond and adapt to a man, by way of conversation, understood in a broad sense (Pask 1971). In a way all competent art is according to Pask implicitly such an environment. This does

not imply that cybernetics requires that we change how we go about creating objects

of art, but it does suggest a novel strategy. According to Pask the conventional art ob-

jects conceal the conversation while cybernetics invites us to externalize it (Pickering 2010). Pask explored the idea of the aesthetically potent environment across a variety

of different focuses – in art, installation projects, interactive teaching systems, theatre, architecture and design tools. Although the physical demarcation between human

and machine in Pask’s different systems is definite, subsystems representing regions of respectively human and machine control are arbitrary and have limits that are con-

tinually changing. This compels the users to assume a role of participant observer. This stance requires that one maximizes interaction, which is in direct opposition to how the traditional scientific observer is expected to minimize interaction (Pickering 2010).

I believe that a number of useful ideas are revealed in this brief entry into the cybernetics of Pask. Firstly:

“If our usual notion of design entails the formulation of a plan which is then

imposed upon matter, the cybernetic approach entailed instead a continuing interaction with materials, human and nonhuman, to explore what might be achieved—what one might call an evolutionary approach to design, that necessarily entailed a degree of respect for the other.” (Pickering 2010, p.32)

Here novelty occurs as a consequence of what I call a decentred agency. Secondly, within this condition, a performative thinking substitutes a representational. According to Pickering “the entire task of cybernetics was to figure out how to get along in a world

that was not enframable, that could not be subjugated to human designs—how to build

129


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

machines and construct systems that could adapt performatively to whatever happened to

come their way“ (Pickering 2010 pp.31-32). In a sense complexity is allowed to remain

complex, by refraining from attempting to produce a reductive representation of it. In-

stead one simply interacts with the complexity and observes and learns from the result. “Man is prone to seek novelty in his environment and, having found a novel

situation, to learn how to control it.” (Pask 1971 p.76) Non-representational technology

The question of representation is important to my understanding of the unique nature and potential of cybernetics, which clearly sets it apart from a conventional scientific

practice. What cybernetics offers is an approach to being in an environment in a non-

representational fashion. We can trace this characteristic back to the very beginning

of cybernetics, where its models immediately appear unusual in a scientific context. Rather than being equations on paper they are performing machines. Pickering notes

that this circumstance exhibits an interesting scientific hybridity between a modern emphasis on representation, and a non-modern attention to performance (Pickering

2010). I believe this same hybridity to be present within the agent-based design space described in the opening of this chapter.

Two examples of such machines were Grey Walter’s Tortoises (1948-1949) and Ross

Ashby’s Homeostat (1948). These were electromechanical devices build with the initial purpose of explicating the workings of the brain as a performative organ. But, what

is interesting in this context is how they, as adaptive machines, suggest and prove that we do not need to construct representations of an environment to act within it – including technologically.

The Tortoises were basically small robots capable of detecting obstacles and light. Following simple rules these electro-mechanical devices were able to move respec-

tively away and towards these elements. Since they periodically emitted light themselves, they also exhibited an ‘interest’ in their own kind resulting in a sort of perpetual

‘mating dance’ where they encounter and loose interest in each other. “The machines cannot escape from one another; but nor can they ever consummate their ‘desire’” (Walter

1953, pp.128-129). Although they acted as models of the brain, more generally the Tortoises illuminated the go of adaptation to an unknown environment. From their

130


A CYBERNETIC APPROACH

[2]

2. Grey Walter’s tortoise | 3. A tortoise in action (agency and environment)

131

[3]


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

behaviour, as simple organisms, we can draw the insight that they did not construct

and process representations of their environment but acted relative to what occurred

in it (Walter 1950). They arrived at solutions without representation. The same applies to the Homeostat. Pickering notes that this machine, “[…] stages for us a vision of the

world in which fluid and dynamic entities evolve together in a decentered fashion, explor-

ing each other’s properties in a performative back-and-forth dance of agency” (Pickering 2010, p.106). It simply consisted of four interconnected units (single homeostats) that converted electrical inputs into electrical outputs (Ashby 1960). These were though

as a means to simulate how essential variables are maintained in an organism (brain). The idea was that the units each maintained a single variable and that they could automatically adapt their configuration, from any initial stage, to achieve stability. Since

a Homestat’s environment consisted of other homeostats it was a responsive and dynamic environment. The Homeostat shows how a solution space can be established

without hierarchy through a decentred negotiation between the elements constituting a system. Both the Tortoises and the Homeostat exemplify a technology, where com-

plex behaviour resulting from the interactions of simple entities can occur without a representational model of an environment.

While the work of Grey Walter and Ashby indicate an interesting relation to rep-

resentation, they do not include the participation of an observer, and their systems are directed towards pre-defined goals. In design-practice goals tend towards being ill-defined, changing and underspecified, and, as discussed above, include the direct

involvement of an observer, as part of the system. To be able to address the nonrepresentational potential of cybernetics we can return to a discussion of the second

order cybernetics of Pask, which can be argued to both extend and depart from work such as Grey Walter’s and Ashby’s. This, through his idea of the aesthetically potent environment and the decentred agency discussed above.

Pask was initially interested in simulating how learning took place – in order to do

so he build physical devices. An example hereof was the project Musicolour (c.1956). Musicolour was a device that used the sound of a musical performance to control a

light show, the aim of which was to achieve a synesthetic combination of sounds and light. Technically the music was converted into an electrical signal. Within Musicolour this signal passed through a set of filters, sensitive to different frequencies, and the

output of the filters controlled different lights. This would have been quite mundane

132


A CYBERNETIC APPROACH

if it had not been for the fact that the internal parameters of Musicolour’s circuitry

were not constant. Analogous to biological neurons, banks of lights would only be

activated if the output from the relevant filter exceeded a certain threshold value, and these thresholds varied in time as charges built up on capacitors according to the development of the performance and the prior behaviour of the machine. Basically

Musicolour was designed to exhibit boredom. What is significant about the project is that one can only learn about its device through performance.

“From the performer’s point of view, training becomes a matter of persuading

the machine to adopt a visual style which fits the mood of his performance. At

this stage […] the performer conceives the machine as an extension of himself, rather than as detached disassociated entity.” (Pask 1971, p.86)

Musicolour staged the encounter of two exceedingly complex systems – a human and a machine. One of Pask’s contemporaries in cybernetics Stafford Beer used this term

to define cybernetics as precisely a science of such systems in which representation

becomes impossible and control is achieved through continuous performance (Beer 1959). Pickering notes that “In contrast to the traditional impulse to dominate aesthetic

media, the Musicolour machine thematized cooperation and revealing […]” (Pickering 2010, p.320) and Pask himself that “[…] the machine which translates between sound

and vision must be a malleable or “learning” device that the performer can “train” (by varying his performance) until it assumes the characteristics of his personally ideal translator”

(Pask and McKinnon-Wood 1965, cited in Pickering 2010, p.320). The interest of Pask was how the introduction of man into the environment of the machine could

supply productive disturbances, thus forming an aesthetically potent environment. This

work continued throughout the life of Pask amongst other in the project Colloquy of Mobiles (1968) (Pask 1971), which was commissioned for an art exhibition, and in a

sense was a sophisticated variant of Walter Grey’s tortoises staging open-ended per-

formative encounters between its human and non-human participants. I will suggest that the significant aspect of these projects is the idea of making control something

we achieve through learning, and that it is the resulting relation between a learning

participant and system that allows a productive difference. I will hypothesise that difference become absent if control is simply given.

Pask’s work with the idea of participation and performance eventually display ties to

133


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

the world of architecture. Work on a proposal for a cybernetic theatre, which could be described as an aesthetical potent environment for actors and audience, led directly to

the involvement of Pask in an actual architectural project – the never realized Fun Pal-

ace by Cedric Price (Mathews 2007). Contemporary with the designs of Archigram, but at a much more pragmatic level, the Fun Palace was intended as a reconfigurable adaptive space that could support an enormous variety of activities that changed over

time. It was to be an architecture that was informal, flexible, un-enclosed, and impermanent. According to Pickering it is impossible to know Pask’s contribution in detail

but: “Pask’s contribution appears to have been to see the Fun Palace on the model of Musico-

lour—as an aesthetically potent environment that in its inner reconfigurations both reacts

to emergent patterns of use and fosters new ones.” (Pickering 2010, pp.367-368). The Fun

Palace project suggests that the cybernetics of Pask extends beyond an applicability as

solely defining a relation between man and machine, but can also be projected as a tool for defining building programs – envisioning the program of a building as an ecology

of activities, events, conflicts and control (Mathews 2007, pp.118-119). This stages the interdisciplinary character of cybernetics in direct relation to architecture, within a non-representational framework.

What I hope to show with the above examples is how cybernetics by contemplat-

ing agency and environment, and by transcending disciplinary boundaries, creates the conditions for a distinctive technological approach to reality. What cybernetics

suggests, across the examples of the Tortoises, the Homeostat, and the devices and projects of Pask, is a technology abandoning representation and hierarchy, creating

a productive environment through an iterative process of performance. Eventually leading to both an architecture and a design process of a non-modern self in which a

continuous process of learning and acting produces a difference in a shared environment between technology and man.

Observers, Reflexivity and Autopoiesis

A key cybernetic concept closely associated with the discussions above is that of reflexivity. If the idea of a decentred agency is understood as a proposed condition of

being, and the idea of an aesthetically potent environment as the ‘ecology’ in which it resides, then the concept of reflexivity can be understood as the necessary operational mechanism by which change occurs. In the words of Hayles reflexivity is:

134


A CYBERNETIC APPROACH

[4]

[5]

4. The Fun Palace by Cedric Price | 5. Gordon Pask on the cover of a publication from the Architectural Association

135


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

“[…] the movement whereby that which has been used to generate a system is

made, through a changed perspective, to become part of the system it generates.” (Also) “[…] reflexivity has subversive effects because it confuses and entangles

the boundaries we impose on the world in order to make sense of that world.” (Hayles 199, p.8-9)

Reflexivity entered cybernetics mostly through discussions regarding ‘the observer’ of

2nd order cybernetics, where the challenge was how to define cybernetic systems so that the observer can be taken into account. A highly influential figure in the early stages of this endeavour was Heinz Von Foerster, who was the first (1960) to propose

models termed ‘second order’ in which the observer was defined as a system itself (Von Foerster 1982). Central to this work was an effort to avoid tying the concept of reflexivity to subjectivity, while retaining autonomy and control.

“Instead of searching for mechanisms in the environment that turn organisms into trivial machines, we have to find the mechanisms within the organisms that enable them to turn their environment into a trivial machine.” (Von Foerster 1982, p.171)

I will take the liberty of proceeding by singling out a mature position elucidating a take on the concept; the autopoietic system proposed by Huberto Maturana and Francisco

Varela. This theory expands the reflexive turn into a fully articulated epistemology that sees the world as a set of informationally closed systems, and extends self-organization to be also autopoietic (self-making) (Hayles 1999). What is interesting with regards to

the Autopoietic system is that it can be shown to relate to the generative position of

Lars Spuybroek in the previous chapter. The design spaces of Spuybroek are systems in which movements within a virtual continuum and a resulting actualization allow the

constitution of concrete architectural morphologies; a similar relationship exists in the autopoietic system. This is acknowledged by Spuybroek himself as a source of inspira-

tion (Spuybroek 2004, p.7). Thus, my interest in the theory of the autopoietic machine follows from its ability to conceptualise a generative approach in a way that connects it to the discussions above, as well as to a broader context of digital architecture.

In ‘Autopoiesis: The Organization of the Living’ Maturana and Varela regard living sys-

tems as unities characterized by autonomy and diversity endowed with the capacity to

136


A CYBERNETIC APPROACH

reproduce, and they seek to understand the organization of living systems in relation

to this unitary character (Maturana Varela, 1980). Herein exists an interest that is not in properties of components, but in processes, and relations between processes realized

through components. Living systems are in this way regarded as machines defined by

their organization and not component qualities. We can here observe a parallel to the topological schema of a generative practice of architecture.

For Maturana and Varela, machines – living or otherwise – are objects defined neither by the nature of their components nor by the purpose that they fulfil. Apart from those

properties within components that participate in the interactions and transformations which constitute the system, the nature of components is irrelevant. Conversely, what is relevant are the relations within the machine and the network of interactions and transformations into which they can enter.

Maturana and Varela distinguish between the organization and the structure of a

machine. The organization specifies the relations that components must generate to

constitute the machine as a unity. Therefore the organization is independent from the properties of components which could be any as long as they produce the correct rela-

tions. The structure of a machine constitutes the relations between components that integrate a concrete machine in a given space. But:

“(…) although a given machine can be realized by many different structures, for

it to constitute a concrete entity in a given space its actual components must be defined by that space, and have the properties which allow them to generate the relations which define it.” (Maturana Varela, 1980, p.77)

Components are in this way related to a context. It could be suggested that the role

played by components in this way resemble that of the material elements within the Morpho-ecological approach suggested by Hensel and Menges; as interface between an interior and its environment, as discussed in the previous chapter. It is not until a

machine is inserted into a context wider than itself that the question of functionality can be addressed. Maturana and Varela adopt the radical view that purpose or aims are not features of the organization of any machine; they are notions that belong in the

domain of descriptions. The organization of a machine only states relations between

components and rules for their interactions and transformations. Thus, the notion of purpose and function has no explanatory value in the phenomenological domain

137


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

which they pretend to illuminate, because they do not refer to processes indeed operating in the generation of any of its phenomena. But any machine, can be describes by

an observer as endowed with a project, a purpose or a function, if properly handled by him with respect to an encompassing context

That living systems are machines can be shown through their organization and the particular properties that arises from their organization. Maturana and Varela term the type of machine that living systems are as autopoietic machines.

“An autopoietic machine is a machine organized (defined as a unity) as a net-

work of processes of production (transformation and destruction) of components

that produces the components which: (i) through their interactions and transfor-

mations continuously regenerate and realize the network of processes (relations) that produced them; and (ii) constitute it (the machine) as concrete unity in the

space in which they (the components) exist by specifying the topological domain of its realization as such a network� (Maturana Varela, 1980, p.78-79)

A continuous change occurs within an autopoietic machine specified by the network of processes of production of components which constitute its organization. An autopoietic organization is a closed domain of relations. It defines a space in which it

can be realized as a concrete system; a space whose dimensions are the relations of production of the components that realize it. If stopped for observation we find spatial relations among its components, but the observed spatial relations do not and cannot

define it as autopoietic. In contrast a crystal is defined by the spatial relations that define the relative position of its components.

Autopoietic machines have an individuality that they actively maintain independently of their interactions with an observer. Their operations specify their own boundaries

in the process of self-production. This means that they subordinate all changes to the

maintenance of their own organization with disregard to how this otherwise transforms the structure of the machine. Other machines have as the product of their functioning something different from themselves and have boundaries that are defined by an observer, who by specifying its input and output surfaces, specify what pertains to it

in its operations. Therefore these machines can be referred to as allopoietic (allo: from Greek – different, other, another).

Autopoietic machines do not have input or output. They can be affected though, but

only through structural changes and these changes are still subordinate to the main-

138


A CYBERNETIC APPROACH

tenance of its organization. This allows us to put them in our domain of manipulation

and to observe them in the context of a domain of interactions, which is external to their organization. Hereby we can interpret perturbing events as input and the compensation to these perturbations as output.

When interacting with an autopoietic system we project the system unto the space of

our manipulation. We can do this because we interact with the components through

the properties of their constituting elements that are not part of the autopoietic space. We modify the structure by modifying the components.

Autopoiesis is a radical shift because it changes the idea of the informational feedback loop, since this no longer functions to connect a system to its environment.

“In the autopoietic view, no information crosses the boundary separating the

system from its environment. We do not see a world ‘out there’ that exists apart

from us. Rather we see only what our systemic organization allows us to see. The environment merely triggers changes determined by the system’s own structural properties. Thus the center of interest for autopoiesis shifts from the cybernetics of

the observed system to the cybernetics of the observer.” (Hayles 1999, pp.10-11) Reflexivity in this connection is constructed on positionality rather than personality.

This is an interesting move because it changes what the opposite to objectivity is. Thus, rather than being subjectivity, the opposite of objectivity is relativism.

When I began this account of the autopoietic system I noted that my interest in

the theory follows from its ability to conceptualise a generative approach in a way that connects it to the preceding observations made within this chapter. A genera-

tive system is (reminiscent of ) a living system – not mimetically but operationally. As a theory of such systems, autopoiesis designates a relation to context, a notion of function and the role of the observer, in this case the designer. This role is consistent with what has been described already and so makes it possible to think a generative process more directly in relation to a condition of decentred agency, the ecology of an aesthetically potent environment and clearly also a non-representational performative

practice. What I will suggest is that the theory of autopoiesis allows a point of inter-

face between these concepts and the topological schemas and actualised morphologies of a generative architectural design practice.

139


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

Synthesising with Cybernetics II

While the first part of this chapter sought to connect experiences from my own prac-

tice to the field of cybernetics, this second part can be seen to also indicate a relation between the findings of the present chapter and certain aspects of the digital practices described in chapter four. The main point of interface is the ubiquitous concept of agency within these practices.

In the previous chapter I described the complex information models of a contempo-

rary digital practice of architecture as representing a notion of design in the form of

a system which extends beyond disciplinary division, in time as well as space. Here, to design is to activate the agency of the model – to steer relative to information. Thus, the agency of the model is contingent on the agency of the designer (and other collaborators). Consequently I suggest that the main condition characterizing all participating systems, within such a contemporary design space, is that of a decentered agency. Precisely as this is also the case in the Lamella Flock project.

In the previous chapter I described certain tendencies related to generative theories

of design, where the material assume agency and act as interface between design and environment . Here we find an approach to design that repositions evaluation of material characteristics and behavior away from a rigid description, towards one that pro-

motes responsiveness. That is, opposing an absolute form of representation and instead subscribing to a performative understanding of design practice. I will suggest that Pask’s concept of an aesthetically potent environment conceptualizes this condition

in a fashion adequate for counteracting the danger of legitimizing this performative perspective through a notion of optimization – fashioning it in the form of an instru-

mental reason. What becomes the main concern here is performance as a mechanism capable of producing a difference in a shared environment between technology, material and designer.

In the previous chapter I described the generative as a specific approach to design

that orchestrates formation by means of design procedures, originating in topological diagrams. As already noted the idea of operating by means of a topological thinking

towards establishing concrete architectural morphologies resonates with the model of

the autopoietic machine. The organization of the autopoietic machine corresponds to the topological diagram of a generative process – the structure and components matches the actualized design proposal. I will suggest that the autopoietic system can

140


A CYBERNETIC APPROACH

act as a conceptual model that establishes a relation between the generative as approach and the mechanisms of performance, agency and reflexivity as described here.

By relating the findings to the tendencies of a general context of digital architecture, identified in the previous chapter, I also mean to suggest that they flow in the direction

of the thesis’s framework for design. That is, the framework aligns to these tendencies. For this reason I will close this chapter by offering a provisional description of the design spaces native to the framework. Thus, these design spaces are systems characterised by:

1. A condition of decentred agency amongst their constituents – technological, material, digital, human (disciplinary)

2. Being environments that can be approach in a non-representational and non-hier-

archical fashion – allowing complexity to remain complex

3. Operating by means of the mechanism of reflexivity staging a complex interaction

between auto- and allopoietic machines – far from a mechanistic instrumental reason of input/output equations

In the next part of the thesis theses characteristic will reappear in different forms, and will be additionally expanded upon and appropriated.

141


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

[6]

6. & 7. A three-dimensional cellular automata (by author). The system created as a probe investigating the use and abstraction of natural complex systems as an approach to the development of form. Cellular Automata (CA) is a nearly canonical form of complex system. CA are a type of computing machine, which is a dynamical system that is discrete in both space and time. The formalism for cellular automata was invented by John von Neumann (with help from Stanislaw Ulam) in the 1940s as a framework in which to study the process of reproduction. In this case the essence of reproduction is abstracted away from how animals reproduce. Instead von Neuman concentrated on the simplest mathematical framework that would allow information to reproduce (Flake 1998).

142


A CYBERNETIC APPROACH

[7]

143


PART 2



EXPERIMENT 1 | LAMELLA FLOCK


LAMELLA FLOCK

The Lamella Flock project introduced the idea that an agent-based solution space can

be used to establish a design practice characterized by the simulated agency of material

components, under the ‘steersmanship’ of the designer. As described in chapter five, through this the project revealed a mode of engaging reality, architecturally, that I had not previously been explicitly aware of. The project’s point of departure predates the

commencement of the thesis – as such it poses questions outside of this context – but was concluded well into the first year of the project. In this role of transitional experiment, findings, and processes leading up to these, had a substantial effect on the outset of the thesis. However, in the context of this text I will emphasize how the project

supplies an understanding of a potential relationship amongst current tendencies of a

digital architecture. Namely, on the one hand, an informational consistency between design and realization, and on the other hand, a generative approach to design – both of which has been sketched out in chapter four.

The mode by which the project engages the design process has already been presented. Nevertheless, I will start by offering an example that illuminates the understanding of design suggested by Lamella Flock.

Andrew Pickering proposes two currently prevailing ontologies. He exemplifies these

respectively by the painting practices of the late Piet Mondrian and Willem de Koon-

ing, which he reads as ‘philosophical objects’ (Pickering 2008). Mondrian’s geometrical abstracts constitute a detachment from the world. They are contemplated and depart

from within the mind of the artist; a practice dominating the world from the outside as an autonomous agency in a passive material world. While being just as abstract, De Koonings paintings cannot be imagined as the translation of a preconceived mental

image translated into paint on canvas. Rather they represent a circular movement between perceptions of emergent effects in the material of the paint and attempts to

improve upon these same effects. Thus, Mondrian’s work represents a dualist movement of human detachment from the world, while de Kooning’s represents a constitutive engagement with it.

When Pickering elaborates further on de Koonings practice I believe that he con-

jures a set of characteristics with similarities to how Lamella Flock contributed to the understanding of design pursued within this thesis. De Koonings work represents a

symmetric interplay between human and non-human, i.e. an observer and the material of the paint, and it thematises time, since it can only evolve in the real time of embodied practice (Pickering 2008, p.2). As will be described below Lamella Flock con-

147


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

[1]

[2]

[3]

[4]

1. ‘Broadway Boogie Woogie’ (1943) by Piet Mondrian | 2. ‘Black Untitled’ (1948) by Willem de Kooning | 3. Parametric Wood Construction (2007) by Martin Tamke (principal researcher) & Jacob Riiber | 4 The original Zollinger construction

148


LAMELLA FLOCK

structs a design space existing exactly between a human and nonhuman agency, and where the artefacts produced, as with the paintings of de Kooning, cannot be given in advance, but must evolve in time – within an ontology of becoming.

However, the design space constructed by the Lamella Flock project was not itself the aim of the experiment. Rather, it developed as a solution to a problem related to the

construction and design of a material artefact. The project was initialized by associate professor of architecture Martin Tamke (CITA) with my assistance. Our outset was

to investigate how a specific ‘traditional’ wood construction system – the ‘Zollinger’ structure – could be transformed to accommodate a non-standard practice, with re-

gards to design as well as realization. We wanted to establish a digital platform that would allow us to design and fabricate freeform structures based on the principles of

the Zollinger. This was part of a larger research trajectory engaged in intersections be-

tween parametric modelling, mass-customization and the use of wood beam elements. In this specific instance the dominant challenge related to managing a design process in light of the concept of circular dependencies. This concept was introduced into the

project in two ways; firstly, by way of the pattern of the wood construction in question, secondly, by way of the design process itself which sought to establish a non-hierarchical informational consistency between fabrication and early form-development.

Preliminaries

The Zollinger construction is a type of lamella roof construction invented in the 1920s as a means to create wide-spanning structures out of short pieces of timber (Allen

1999). The system’s structural principle consists of a crisscrossing pattern of parallel arches comprised of relatively short members. These are hinged together in a diamond

pattern forming an interlocking network. The ingenuity of the system resides in two properties: the efficient joint system that minimizes the amount of shared meeting points allowing for simple assembly, and structural strength given by the interwoven beams. Historically the system manifests itself as roofs with barrel shapes, while

similar constructions, such as reciprocal frame systems (Popovich 2008), usually form domes. This naturally limits their formal adaptability. We observed that this condi-

tion must originate in the complexity of dealing with the interdependencies between its elements. Dome and barrel shapes are the product of relatively simple procedures

within the system, while our desire to produce alternative shapes constitutes consider-

149


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

ably more complexity in development and design. What represents a challenge in these circumstances is that any change made to a single beam element will alter all other

beams in the structure as well – not only the immediate neighbours, since everything is reciprocally related. No existing solution could be found in this regard, unless one

abandoned the concept of using straight beam elements with precisely positioned joints. Here work from the Architectural Association School of Architecture (Hensel Menges 2007) and Shigeru Ban (Tristan et al 2007) demonstrates a principal ability of similar

systems to form different shapes using the flex of the material, tolerances in the joint geometries, and changes in the system’s local orientation. Within these projects we noticed

that design solutions were achieved by following a bottom-up approach, where each ele-

ment is treated individually as it acts autonomously in a larger formation. This was a key observation used to guide us towards the final solution for a digital platform that would allow the design of freeform structures.

The difficulties attached to the circular dependencies within the formal setup of the

structure, were combined with the challenge of producing solutions that would be realizable as 1:1 structural constructions in wood. Previous research on mass customized parametric wood constructions (Tamke et al. 2008) indicated that a digitally controlled pro-

duction can provide a flexible, effective fabrication of easily assembled wood beams. This approach is based on the conjunction of computation, digital fabrication, and traditional

craft techniques, where current CNC wood-joinery machinery allows for the cutting of joints that are adaptable to a variable geometry, at high speed. These joints allow for fast

assembly as they incorporate self-registering geometrical properties (Schindler 2009). Here, new applications of a knowledge related to traditional wood crafts are qualified in combination with an improved understanding of forces within massive wood through the use of structural simulation systems (Holzner 1999). The blend of computational

capabilities and digital fabrication allows craft related knowledge – previously bound to the skill and knowledge of the executing craftsperson – to be introduced into a contemporary digital practice.

This created another level of circularity within the project – between model, structure,

material and fabrication – which needed to be implemented within our design-system, creating a feedback between designing the structure and its future as a physical full scale artefact. In this, the outset of the project related more to the tendency within digital ar-

chitecture concerned with an informational consistency between design and realization, than that of the generative appearance the project would eventually also have.

150


LAMELLA FLOCK

[5]

[6]

Straight elements

Reciprocity

Freeform

Angle limitations

Offset point of connection

Size limitations

5. The reciprocal pattern of the Zollinger construction | 6. Restrictions imposed by the system and its production

151


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

Project Execution Early Probes

In the initial stages of the research, the distribution and possible computing of el-

ements were investigated, looking for the most suitable method of controlling the system and the non-linear relationships within it. Here two failed preliminary tests

exemplify the challenges inherent within the system, and show how we were directed towards the final agent-based solution. In the first test, we attempted to ‘force’ the

Zollinger structure onto preconceived freeform surfaces, by projection of its pattern. This top-down approach quickly proved unsuccessful, due to the impossibility of placing the straight beams correctly on a curved surface. This resides in the fact that both

beam endpoints and midpoints, which connect to endpoints of other beams, need to

intersect with the surface. Broadly speaking the method fails because no local nego-

tiation is possible between structure and surface. In the second test, we attempted a bottom-up approach instead. Already in early studies this immediately proved more promising, but presented significant problems regarding design control. In this test we

tried to mimic how one would manually go about crafting the structure by combining elements one after the other. This resulted in a computer model that distributed the

structure sequentially. Results produced by this model complied with the principles of

the system but restricted design control. The cause of this was that the reciprocities within the structure came to be directed by the sequential distribution of elements

in such a way that any manipulation made would propagate through the structure with an exponentially increased effect. This resulted in interesting but unpredictable configurations of the overall form of the structure, which limited the model’s use as a design tool.

Although unsuccessful, these two initial test-approaches allowed us to state the re-

quirements of our system and propose a solution. Firstly, we learned that form could not be imposed from outside the system and required a degree of negotiation; sec-

ondly, we learned that linearity resulted in mechanisms that impeded our ability to act as designers. Thus, we needed a bottom-up process with the ability of dynamic non-

linear interaction, where different design possibilities could be explored. This led us to introduce an understanding of the structure as a self-organizing system consisting of entities possessing a simple set of behavioural properties and relations to each other.

152


LAMELLA FLOCK

[7]

7. Construction made through early prototype parametric system

153


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

The System: principles, structure and behaviour

We arrived at this understanding by encountering the idea that a group of agents, comprising a self-organising system, may be able to perform tasks without explicit representations of their environment or other agents, and the idea that planning in

this context may be replaced by reactivity (Carranza Coates 2000). This prevents the

problems associated with the sequential distribution of elements above, in that a bottom-up approach here is free from the planning ‘ahead’ of a sequential order. The ad-

vantages of using self-organization to solve problems reside in a flexibility to function within changing environments and an ability to function even though some entities

may fail to perform. The disadvantages can be located in the bottom-up approach to specifying such systems. Here paths to problem solving can never be predefined but

are always emergent and result from interactions among entities themselves, as well

as between entities and their environment. Therefore, using self-organization to solve a problem requires precise knowledge of both the individual behaviour of agents and

what interactions are needed to produce a desired global effect (Bonabeau et al. 1999). The conceptual move towards conceiving the design system by means of principles of

self-organization required a change to our existing digital practice, which at this point mainly departed from parametric modelling software combined with the occasional

use of scripts. Instead our digital model necessitated that the system was developed by writing an actual computer program capable of simulating the reciprocities aimed at. This program departs precisely from specifying the behaviour required by the components within the system. Representing parts of the Zollinger structure we conceptualized these components to be organizationally distinguished as four line segments

unified in a spiralling motion of mutual support. In this way each entity exhibits, within itself, the non-linear relationship that also defines the global structure aimed at. The Program

To initialize the program consists in determining the amount of components, their

sizes, and a preliminary dispersal of these in space, e.g. their initial condition. While running, the system is controlled through four behavioural algorithms that accumulate vector information describing the trajectories of components as they, in time, attempt

to form a stable whole. A technique inspired by how the identification and replication of different goal types can be used in the simulation of flocking behaviour (Flake

154


LAMELLA FLOCK

[8]

8. Inspiration: Flocking behaviour

155


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

1998). Each algorithm produces directions and velocities that interact to produce the

overall movement and transformation of an entity. The algorithms are named and described thus:

Move: If not representing a corner or an edge, each entity has four neighbours. By measuring the distance and direction from unattached endpoints of beam representations to a neighbouring connection point, vectors are calculated. These vectors – often conflicting – are added and weighted to calculate a mean vector by which the component is moved.

Orient: By altering the configuration of angles between beam segments, each element tries to orient its segments towards their neighbours. A seg-

ment is in this way sought to be aligned with the trajectory towards its destination.

Stretch: Through the above orientation, a segment will, within a certain

tolerance, be able to stretch to connect to a neighbour. This is allowed

when the orientation is correctly aligned, and if it is happening within a predefined limitation to size determined by the fabrication technology.

Scale: Each component has the ability to increase and decrease size while

keeping its proportions. This allows for a global ‘push/pull’ effect within the lamella network.

Implementing Production Constraints

Additionally production related constraints were introduced into the program. At the

scale of the individual beam elements this means that the computation restricts the beam sizes and angles of intersections to the specifications allowed by the machin-

ery used for production. Also, the program is informed by the fact that two beams

cannot share the same meeting point on a third beam. That is, even though all nonedge beams will be connected to two other beams midway along their length, drilling

for joints needs separate space for both connections. The program deals with this by slightly offsetting the shared meeting points of every element away from each other.

The generative design process was in this way informed by its implementation and realization in 1:1.

156


LAMELLA FLOCK

Move

Initial Stage

Scale

5

5

4

4 1

0

1

0

6

5

6

5

7

4

7

4

2

2

6

6

3

3

7

7

Orient

Stretch

5

4 4

5

1

1

0 6

5

6

5

7

4

7

4

2

2 6

6 3

3

7 7

[9]

9. The behavioural principles

157


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

[10]

[11]

10. Global behaviour: Negotiation of the agents | 11. Stabile configurations of the system

158


LAMELLA FLOCK

Behaviour and Method

The global behaviour occurring from the algorithms of the program produces a network of entities that attempts to obtain a shape describing a coherent surface. The

global configuration is continuously and non-lineally renegotiated until a stable result

is achieved. To the designer observing this, the components appear as simple organ-

isms struggling to comply with the goal of finding a stable state by means of the

behaviour they have been granted. As described in chapter five, from the perspective

of developing the system, it was a matter of observing this behaviour in time, in order to correct the algorithms until results were achieved. That is, iteratively rewriting code

and observing the resulting process. This clearly breaks fundamentally with conven-

tional architectural ‘drawing’ techniques, although the representation of an architec-

tural artefact is still the final purpose of the endeavour. I will return to this question of representation and agency in chapter six.

In keeping with the role of the designer, previous experience (Tamke Thomsen 2009)

shows that in the context of architectural design a combination of a generative logic and human agency has an augmenting effect on the capabilities of both. In Lamella

Flock we introduce the possibility of actively manipulating entities while the system is running, thus forming a feedback loop with the designer, rather than positioning

him/her in the role of initializing the process only. This results in a tool where changes in the configuration of a surface can be made by altering local conditions, while selforganization deals with the global consequences of these actions.

Actions available to the designer include the ability to move an entity in any direction

or change its size, as well as fixate it in a given position. This last feature forces the surrounding network to adapt to the new conditions. Colour coding of elements and

a navigational diagram helps to maintain an overview of these manipulations. Preci-

sion and localization of the design model were given through a millimetre based unit

space and the ability to import 3D representations of a site to which the design could be aligned.

Designing Material Evidence

The self-organizing system of Lamella Flock allows the model to interact dynamically

with and inside an environment given by material, structural logic, design intend and production. Changes to the environment through manipulation are instantly answered

159


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

by the model through dynamic shape change. These transformations appear to the designer as a result of an internal reflection rather than direct answer. As described

in chapter five this produces a situation where designing with the system, similarly to

the process of designing the system itself, proceeds by a process of learning about the distinct character of the model and its behaviour.

In the process of reaching a final design proposal the output was adjusted to different model scales ranging from design speculation to 1:1 realization. Thorough dialog and

testing through prototypes were crucial to determine the adequate types and dimen-

sions of joints, fasteners, bearing and bracing for fabrication and assembly strategies. Feedback from findings related to these prototypes was iteratively integrated into the

model. The incoming information was handled in a pragmatic way where new insights were either encoded as internal conditions in the generative code or the visual interface was used for constraining the self-organizing system appropriately.

This preparation allowed us to exploit the capacity of digital fabrication and self-regis-

tering joinery, demonstrated by only 3.2 hours of cutting time and two days of overall assembly of a 1:1 structure consisting of 80 individualized beam elements.

Reflections

In the opening of this text I observed that the project stages an understanding of de-

sign positioned at the intersection between complex information models – exhibiting

informational consistency across the total process of architecture – and a generative approach to design. The project exemplifies how these tendencies may come together

within an ontology of becoming – not just as a conceptual model, but as a model that serves a concrete purpose of solving formal as well as process-related complexity. To

think the design space as a self-organizing system of interacting agencies suggests a

means to produce a robustness while acting in the context of conflicting requirements, to negotiate form rather than imposing it and to operate non-linearly, while maintaining the necessary degree of control needed to engage in design. In summary this is the contribution I believe that the project makes to the thesis.

The project also introduces a number of problems though – some internal to the proj-

ect, others directed towards the broader context of the thesis. However, I view these problems as productive in light of the present endeavours.

In a sense the Lamella Flock constitutes an automated parametric model, where time

160



GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

[12]

[13]

12. Laser cut model produced from the Lamella Flock system | 13. The Lamella Flock demonstrator at ROM Gallery, Oslo 2010

162


LAMELLA FLOCK

appears to be real and irreversible, but where genuine transformative or unexpected events do not, and cannot, occur. The space in which something singular or different

might appear is greatly diminished. I view this as a result of the degree to which the project requires the specification of constraints and goals. Accordingly, the system

is defined by the quantifiable data required to assemble a predefined organizational

pattern of a specific structure and its construction. Adaptability is limited to what

is already known, or what can be approached through manipulations made by the designer. The end result is therefore a system constituting a ‘mono-culture’ of structural components, where compatibility is specified prior to a process of formation, and where the self-similar elements are incapable of interfacing with an unknown outside or the introduction of deviations within the structure itself.

Of course the project should not be interpreted as a proposal for a more general framework for design. That is not its intention. Though, in the present context it does exemplify ideas, techniques and mechanisms that allow me to point towards aspects of such a framework. My criticism is therefore directed towards counteracting difficulties with

regards to such a relation and not the project itself – which proved successful according to its own immediate aim. If the utilization of self-organizing systems – consisting

of interacting component-based agencies – should act as a component within a larger

framework for design, then the problems of a constrained limited adaptability can

be used to state a number of future requirements. Namely, that such systems should be operational relative to an underspecified design space and that they should have the adaptive ability to interface with what is encountered within an unknown or unforeseen environment. Nothing within the Lamella Flock project suggests that the

algorithmic and computational techniques used refuses this kind of versatility. What is required here is to alter the mechanisms observed in the project to operate at a more general level, than was the case.

In the experiments, investigative probes and program sketches I have developed

throughout the present project this search for the underspecified and adaptable have been an important motivation. The illustrations concluding this chapter stems from a minor experiment by which I explored the idea of ‘growing’ patterns following the

reciprocal logic of the Zollinger. However, by following a more locally defined and general principle, I was able to develop patterns in which absence, noise and difference do not prevent the system from arriving at a solution.

163


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

164


LAMELLA FLOCK

[14]

14. Patterns for reciprocally supporting beam structures generated so that absence, noise and difference do not prevent the system from arriving at a solution. Examples and generative principles

165


6 | REPRESENTATION AND AGENCY


REPRESENTATION AND AGENCY

While previous chapters have overflown the territory of the investigation in order to delineate, separate and make appropriate selections, what follows can be understood as the beginning of a consolidation of the material, from within this territory.

In this chapter my aim is to address the topic of architectural representation with re-

gards to the framework of the thesis so far. Consequently, insights from the preceding

chapters will naturally have implications for how one might proceed in this matter. In relation to the nature of the thesis as a whole, the research context itself implies a

reimagining of the concept of representation. How does one sustain an operational overview within a highly fluid framework for design, which mostly departs from nonstandard practices of modeling? Practices that are as much algorithmic as they are

geometric, material or visual. The present practice – as well as any critical practice departing from digital technology – will eventually have to address the question of repre-

sentation. As outlined in chapter four this digital field is characterized by introducing

elements that are foreign to conventional drawing practices – topological diagrams, material performance, interdisciplinary shared models etc. These elements expand

upon the already technologically saturated concept of representation in architecture. Because of this coupling between drawing and technology the findings in chapter three are as well of significance. If we are to address the ideological implications out-

lined here representation and its technologies should first and foremost be qualified

by seeking out difference and a non-reductionist approach to the surrounding world. Lastly, through the practice of cybernetics chapter five introduced performance as a

contrasting concept to that of representation. While representation by definition seeks to represent nature, and thus produce knowledge that maps, mirrors and corresponds to its object, performativity implies something else. Rather than thinking of the world

as consisting of facts and observations, one can proceed by the idea that it is in fact filled with agency (Pickering 1995, pp.5-6). This presence of agency is already observ-

able in the contemporary field of digital architecture described earlier. However I will maintain that forms of representation are still a necessity to the field of architecture –

if nothing else they serve a practical, evaluative and communicative purpose. Thus, the question becomes how one might position the making of models and drawing relative to the more fleeting reality of agency?

My goal in this chapter is to conceptualize agency as a mechanism introduced into representation – a mechanism that allows representation to assume a reflexive relation

167


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

to that which is represented. In this way I attempt to refrain from subscribing to an

oppositional schema, by thinking representation as a potential vehicle of agency. The discussion departs from certain considerations related to architectural representation

in general. What is of interest here is how certain problematic characteristic and as-

sumptions, attached to conventional drawing practices, potentially persist within new digital forms of ‘drawing’. The main concern here is how complexity is approached and how this easily leads to an unfavourable reductionism. This directs the discus-

sion towards addressing the nature of how contemporary complexity can be perceived. Here the concept of becoming, and implicitly non-linear temporality, plays a major part. After a discussion of these matters, I will in conclusion return to the concept of

representation and present proposals as to how performativity and agency can start to sieve into our models for design.

Representation

The making of architectural representations have often been thought of as an attempt

at maximum preservation in which meaning and likeness are transported from idea to building, with minimum loss. This characterization was problematized by Robin

Evans, who in his essay ’Translations from Drawing to Building’ critically reflects on the role of the conventional architectural drawing (Evans 1997a). I believe that Evan’s concept of drawing and the observations made, with advantage can be used to reflect on questions of digital forms of representation as well.

In Translations from Drawing to Building Evans observes that, in general, architects never work directly with the object of their thought but through an intervening me-

dium. The activity is therefore never the construction and manipulation of the final

artefact, but to try giving a complete determination of it in advance. This builds on the assumption that no information is lost in the process of translation from drawing

to building, as well as a supposed ability to gather all relevant information within the drawing. These beliefs have clear commonalities with the ideal of the optimized model discussed in chapter three, where the act of optimisation is necessarily dependant on

an ability to predict. But there are certain relations, properties and qualities that cannot be expressed in a direct sense in concrete visual form. This is the reason why

standard modes of drawing, according to Evans, should be abandoned. In the essay

The Developed Surface he proposes: “[…] to understand architectural drawing as some-

thing that defines the things it transmits. It is not a neutral vehicle transporting conceptions

168


REPRESENTATION AND AGENCY

into objects, but a medium that carries and distributes information in a particular mode. It

does not necessarily dominate but always interacts with what it represents” (Evans 1997b,

p.199) - in the vocabulary of the thesis, this begins to indicate a condition of agency. The transitive, communicative properties of the drawing could be used to better effect

and Evans wishes for a situation where the corporeal properties of things made, are combined with the more intangible features of a design. Evans shows that contrary to artistic drawing, where the subject is held to exist prior to its representation, within architecture the subject-matter will exist after the drawing. In this way a conventional

construction of representations in architecture is never done relative to material real-

ity but prior to construction. However, as Evans notes, we can never be quite certain, before the act of realization, how things will be transported and what will transform them on the way (Evans 1997a). We may though try to take advantage of the situation

by extending the journey, maintaining sufficient control in transit to evolve the potentialities that might be brought into existence through a given medium.

We can notice similar concerns in the position of Stan Allen, who offers a definition

of the architectural drawing as: “[…] an assemblage of spatial and material notations that can be decoded, according to a series of shared conventions, in order to effect a transformation

of reality at a distance from the author” (Allen 2009, p.41). Allen shares with Evans the

view that the link between drawing and material reality is complex, changeable and impure. What Allen proposes is that we do not attempt to solely follow an eidetic pathway from representation to build artefact, but also one of transposition operat-

ing by way of notation. Many aspect of architecture cannot be represented. However, notation, which makes no attempt to approach reality through resemblance, is capable

of anticipating the complexity and unpredictability of the real. In this Allen points towards a double-sidedness of architectural practice, stressing that drawing must at one and the same time be concrete and abstract. Here he finds inspiration in the

work of the philosopher Nelson Goodman who distinguishes between two types of art forms, autographic and allographic (Goodman 1976). The autographic arts depend on the direct contact of the author (e.g. painting) while the allographic depend on

reproduction at a distance from the author (e.g. musical composition). Architecture is

special because it is necessarily always conducted as a mixture of allo- and autographic activities. Thus, architectural design forms a system that is at once highly abstracted

and self-referential, while at the same time it aims at instrumental transformations of existing reality.

169


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

Within a digital practice we might easily be led to believe that with the arrival of parametric digital modelling (see Woodbury 2010) we become better equipped to

solve the challenges observed by Evans and Allen. As described in chapter four, today we are able to create an unprecedented informational consistency within these mod-

els offering the ability to embed information regarding the realization of a design, within the representation of the design itself. This could, amongst others, be construction constraints linked to digital fabrication, building program, site restrictions or the material behaviour of a structure (Tamke et al. 2008). Does this then achieve the extension of the journey between drawing and building that Evans seeks and the

merge between notations and the eidetic of Allen? I will propose that it does not au-

tomatically do so, and that it might even assume a form where it acts against design as the discovery of novelty. This depends very much on how one conceptualizes, or

extends, a so called parametric model. If we begin by assuming the most simplified

version, I will suggest that the act of parameterizing a design easily introduces a new form of absolute representation. This occurs when it is believed that design develop

by way of stating the necessary constraints that the parametric model will solve, i.e. prediction. This is a standard assumption observed in some discourses of parametric modelling (Woodbury 2010), which I view as reminiscent of the analytics of the Design Methods movement. Here design departs from descriptions of fixed or limited

ranges of possible relationships, and transformations constrained by limiting values. Models come to appear plastic, but relative only to an internal organization, which does not evolve beyond variations in which this organization persists. Yet again, the journey from design to build artefact is contracted to prevent noise and disturbances from an unruly material.

The main problem occurs in refusing to acknowledge that our reality consists of systems that are so extensive and complex that they cannot be described through absolute representations. This includes ones based on variable ‘parameters’. Instead the parameterized model is in danger of becoming a formalism replicating an image of

an idealized timeless nature, where factors of importance have been carefully selected, quantified and isolated. In this way it becomes a kind of positivist image that, although

arguably instrumental to some practical end, is comparable to the problems attributed to any aesthetic formalism. That is, in both cases a pre-determined image forces the

direction of the representation as it progresses into reality; transmitting, but never

170


REPRESENTATION AND AGENCY

receiving, representing, but never performing. This process should rather be characterized by a reciprocal relationship between the ranges of agencies responsible for defin-

ing a design solution. If we look towards the natural sciences professor of philosophy

Nancy Cartwright argues that the fundamental principles of theories of physics do not represent what happens in a material system, they only describe abstract relations

between abstract concepts, “…it tells us the ’capacities’ or ‘tendencies’ of systems that fall under these concepts. No specific behavior is fixed until those systems are located in very spe-

cific kinds of situations. When we want to represent what happens in these situations we will need to go beyond theory and build a model […].” (Cartwright 1999, p.242). Similarly, in

design the parameterizing of representation can be proposed to become a representa-

tion of theory, rather than what actually happens or is required beyond representation. Naturally architecture will always be coupled to representation. However – paraphras-

ing Evans and Allen – a coupling to context, notationally relating representation and

materialization, could serve as a tool to establish a productive, rather than restrictive, kind of prediction, having more to do with simulation, than determinism.

In order to achieve this, a minimum requirement of our digital models would therefore be that time should exist as something real within them. If we recall the distinction

between a reversible mechanistic time and irreversible vitalist time (discussed in re-

lation to the position of Scott Lash in chapter 3) then the parameterized model, as

an absolute representation, assumes the first. If the range of values and relationships within the model are fixed we can retrace our steps, consequently rendering time as

reversible. We can always revert back to previous forms of the model. Memory or ex-

perience is never accumulated within these models, and they are in fact not as dynamic as it might seem on the surface. They lack tension and adaptability, because they make everything internal to themselves. They assume a linear causality far from the vitalist

and generative self-organization that Lash proposes to be characteristic of contemporary culture. This assumption, on the part of the parametric model, imposes in a sense

a way of thinking on the design and therefor also a certain practice; an unjustifiable pragmatism. If another situation is sought, it is then possibly a matter of contemplat-

ing how one could think differently, and to subordinate architectural representation under the result.

171


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

From Being to Becoming “The traditional concept is that things are in time, whereas the new concept is that time is in the things”- Karlheinz Stockhausen1

In summary there seems to be two concerns at play in the above. Firstly, we can observe

a problem of absolute representation and assumed linearity. Secondly, we encounter a problem of reductionism. These are instrumental in establishing the challenges facing an attempt to establish a productive ‘seamlessness’ between representations and acts of

making. In order to arrive at another point of departure for the notion of representation, I believe that it is useful to look towards opposite notions to those of the problems; the interim, non-linear and complex. Here contemporary ontological considerations abandoning a notion of being in favour of the concept of becoming are of special

significance. There are two fundamental assumptions attached to these theories: that time and space are not separate, and that the genesis of form can be accounted for without the aid of a transcendent cause. In the following I will sketch out the basics

of what is at play in these positions, to subsequently readdress the question of representation in light of this theoretical foundation. It should be noted that the theories

in question are not observed as models that have a necessary relation to digital design

technology. Moreover, the tools and techniques of a contemporary digital practice by far exceed what these theories, as a conceptual model, prescribes. However, they supply two things in relation to the present endeavour. Firstly, addressing theories of becoming allow me to arrive at a conceptual and operative vocabulary appropriate for discussing

the nature of the journey implied by Evans and Allen; the space in-between drawing and building. Secondly, later in this chapter I will use this conceptualisation to assign a specific tactical model to practice, through which I suggest a cartographic nature for what might constitute architectural representation in the context of the thesis.

Classical dynamics in physics makes no distinction between the future and the past. The world, as presented here is invariant with respect to the direction of time. Un-

derstood in the sense that, if we reverse the sequential order of a series of events, comprising a process, the relevant properties of the process do not change. Already 1 Quote borrowed from the 2nd issue of Manifold, a Rice University based journal of architectural theory

172


REPRESENTATION AND AGENCY

in the beginning of the 20th century Henri Bergson observes that classical physics describes a reality where change is nothing but the denial of becoming and time is only a parameter, unaffected by the transformation it describes; where dynamics seem

to form a closed universal whole (Bergson 1998). This is the image of a stable world where the process of becoming has been left out of the equation. However, throughout

the 20th century science became aware that instead of finding stability and harmony,

as supposed in classical physics, we find evolutionary processes leading to diversification and increased complexity (Prigogine 1980). Moreover, classical physics assumed that initial conditions of a physical process could be exactly stated. Instead it turned

out that we know the initial conditions, even in simple dynamical motion, only by ap-

proximation, whereby future states of motion become more difficult to predict as time increases (Prigogine 1980). In effect, should time in physical processes be reversible the notion of becoming would be an illusion, and in a sense, time as well.

Turning towards the theory of architecture, in his book ‘Architectures of Time’, Sanford

Kwinter seeks to implement time as something real; as part of a field expressed as the complete immanence of forces and events. This field dispenses entirely with the

need to posit a material substratum as carrier of forces and events. Instead it describes

a space of propagation of effects, distributed through functions, vectors and speeds

(Kwinter 2001). Time and space are here no longer separated as three dimensions of space and one dimension of time, but rather as a four dimensional continuum irreducible to its component dimensions. Matter is here drawn into a process of becomingever-different, and it is hypothesised that it is only through such a process that novelty can emerge.

Morphogenesis

Central to the argument of Kwinter is the definition of two types of morphogen-

esis; i.e. the emergence and evolution of form. According to Kwinter the question of

morphogenesis is at the heart of all formal aesthetic practice, and design practice in particular.

In a classical conception of morphogenesis an eidetic pathway is followed from the

possible to the real; where an already formed image is realized. This is of course ana-

logue to how Evans observes conventional drawing practices. In opposition to this, Kwinter states that an alternative exists, where morphogenesis follows the dynamic

and uncertain process that characterizes the schema linking a virtual component to an

173


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

actual. The difference, between the two types of morphogenesis, is that the possible has no existence, whereas the virtual does; and the virtual does not resemble the actual as something preformed or pre-existing. The relation between the virtual and the actual is then one of difference rather than resemblance.

“Realization (of a possible) and creation (through actualization-differentia-

tion) are two intrinsically distinct and irreducible processes. The first program-

matically reproduces what was already there, formed and given in advance, while the other invents through a continuous, positive, and dynamic process of transmission, differentiation, and evolution” (Kwinter 2001, pp.8-10)

Actualization occurs in time and with time, while realization unfolds the pre-existing

and hereby destroys the possibility for novelty and eradicates time. Therefore the principle of morphogenesis should “be sought in time, within a mobile and dynamic reality riddled with creative instabilities and discontinuities” (Kwinter 2001, p.10). This means

that we cannot conceive of architectural practice at the level of formed objects. Rather the object requires that we acknowledge a situation, where the complexity of reality makes it impossible to engage the world in its totality. This forces us to always operate locally, relative to the unfolding of the world. Objects should here be defined by how

they engage with reality, rather than by how they appear. That is, through performance

and not representation, and through a journey – not of resemblance – but of difference. Multiplicities

In ‘Intensive Science and Virtual Philosophy’, an account of the science that forms the basis of the ontology of Gilles Deleuze, Manuel Delanda presents a similar notion of

the object (Delanda 2005). Here the result is a situation where the complexity of the

object is replaced by the complexity of the space in which the object evolves. This can

occur by making this space immanent to the material process, rather than imposing form from an outside. This immanent space originates within the concept of mul-

tiplicity. This concept replaces the classical eidetic perception morphogenesis, based on the notion of essences. While essences are a static concept invented to explain

the notion of resemblance, multiplicities are connected to a dynamic morphogenesis. Identity is in the case of the latter not defined by fundamental traits but rather by the morphogenetic process that gives rise to it. Forms are hereby always historically

174


REPRESENTATION AND AGENCY

[1]

1. Snowflakes: Objects that cannot be accounted for with reference to the concept of essence, but which are the product of the convergence of different flows within an environment.

175


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

constituted entities.

Multiplicities specify the structure of spaces of possibilities that explain the regu-

larities exhibited by morphogenetic processes. This extends the notion of space to be

linked to questions of process, in addition to geometry. This space of possibility links to branches of mathematics, where a space can be defined with a variable number of dimensions and the absence of imposing an extrinsic coordinization. Thus, multiplici-

ties describe a space without exteriority or unity. Kwinter notes that in such a space we encounter a flatness of being, which does not refer to a diminished dimensionality, but rather something akin to the world of fractal geometry; complexity at every scale and

an infinitely variable space between dimensions (Kwinter 2001). This flatness intensifies and increases the importance of individual cells, but at the same time results in an

inability to grasp or form a total picture. All movement done relative to a multiplicity is therefore tactical, aimed at a mapping of a ‘terrain’ from within.

A multiplicity is a space of possible states, where each property of an object is mapped

into one of its dimensions. To study the behaviour of an object is to study a trajec-

tory within the state space as a time based observation. What is discovered by such observation is the appearance of topological features within the state space that act

as basins of attraction for the trajectories. These attractors, or singularities, are topological features that define the long term behaviour of an object. They are the virtual

counterpart of the equilibrium state of a physical system (Delanda 2005). This endows multiplicities with a non-representational relation to the physical artefacts they cre-

ate. Multiplicities give form to process not end product. The same processes realizing the same multiplicity may be physically highly unlike each other. For example, soap

bubbles and salt crystals are both results of minimizing processes – the realization of the same multiplicity.

“Unlike essences which are always abstract and general entities, multiplicities

are concrete universals. That is, concrete sets of attractors (realized as tendencies in physical processes) linked together by bifurcations2 (realized as abrupt transi-

tions in the tendencies of physical processes).” (Delanda 2005 p.21) Delanda proposes a target for a theory of the virtual:

2 Bifurcation relates to the concept of symmetry (the degree by which a property of an object can be left unaffected by transformations applied to the object). It is a so called symmetry breaking event where one singularity is exchanged by another.

176


REPRESENTATION AND AGENCY

“[We] need to conceive a continuum which yields, through progressive differen-

tiation, all the discontinuous individuals that populate the actual world. (…) this virtual continuum cannot be conceived as a single, homogeneous topological

space, but rather as a heterogeneous space made out of a population of multiplici-

ties, each of which is a topological space on its own. The virtual continuum would be, as it were, a space of spaces, with each of its component spaces having the capacity of progressive differentiation.” (Delanda 2005, p.78)

I will suggest that the concept of multiplicities can help us conceptualize the nature of

notation proposed by Allen. A multiplicity gives form to process, not end product, and its ontological status is that of virtuality. The virtual is fully real, as opposed to merely possible, and its reality assumes the form of an underlying diagram. This diagram leaves behind traces of itself in the intensive processes it animates. In chapter three

Scott Lash is shown to refer to intensive processes as algorithmic mechanisms. Simi-

larly I understand notation as possessing an algorithmic character; this, by appearing

as a context-dependant mechanism that has an effect on a design process, while not

resembling the result of this process. Consequently, I will hypothesise that notations, conceptually, are intensives. Kwinter states that it is precisely the intensive processes, and its extensive counterpart, we need to consider if we want to conceive of architectural practice in other terms than uniquely at the level of formed objects. The concepts of intensive and extensive should therefore be clarified. Intensive and Extensive

Delanda presents, as a metaphor for the genesis of spatial structure, the image of a relatively undifferentiated and continuous topological space, undergoing discontinuous transitions that progressively acquire detail until it condenses into the measurable

and divisible metric space, which we inhabit (Delanda 2005, p.56). Here the topological refer to the intensive and the metric to the extensive.

When observed in biology intensive processes direct both the extensive and the quali-

tative properties of organisms. That is – with hints towards the theory of autopoiesis – they define both the spatial structure and the different material, which gives that structure its specific mechanical qualities. If organisms were not individuated in an

intensive environment, their capacity to evolve would be greatly diminished. Accordingly, the intensive here points towards adaption, which is an attribute I will propose

177


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

[2]

2. Biological aggregation. Images of the mature slime-mold amoebae forming a structure

178


REPRESENTATION AND AGENCY

to be a prerequisite in the journey between drawing and building as well. As an example, in assembly line logic things cannot evolve without mutations occurring simultaneously in matching parts, while the adaptive nature of biological assembly does not necessitate this. Parts of an object put together on an assembly line are typically fully

Euclidean. They possess metric rigid properties and are hereby limited in relation to

the kinds of procedures that may be followed from their assembly. By contrast biological assembly of component parts are defined by the topological properties of these

components themselves rather than metric properties. Shape is less important than

connectivity and length less important than connection points. This leads to component parts that are adaptive and exhibit agency.

“Thanks both to diffusive transport, lock-and-key matching assembly, topological

and adaptive parts, on one hand, as well as stimulus-independence, on the other, evolution has an open space in which to carry out its blind search for new forms.“ (Delanda 2005, p.67)

Within architecture Reiser and Umemoto observes that an essentialist approach, equivalent to the assembly line logic, elevates rationality above matter. This prohibits the productive and rich capacity of matter to define and influence form. Allowing this dynamic to operate is a possible way of re-envisioning tectonics and organization (Reiser Umemoto 2006 p.74).

In the study of thermodynamics the intensive and extensive assume a slightly different

role than in biology. The division between them here is that if a quantity of matter is divided in two, its intensive properties are those left unchanged, while the extensive properties will be halved. Since I earlier made a distinction between the intensive, on

the one hand, and the extensive and qualitative (e.g. colour) on the other, this presents a problem, in that indivisibility applies to both qualities and intensities (e.g. colour and

temperature). Instead intensive properties should be considered as something that, in addition to being indivisible, averages; as when two objects with different tem-

peratures meet. This supplies intensive properties with a dynamical aspect not shared by qualities. Intensive differences are positive and productive; they drive processes of

individuation. They control fluxes of matter and energy. This links the intensive to what Ilya Prigogine terms the physics of becoming, where time ceases to be reversible (Prigogine 1980).

179


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

Naturally, thermodynamics and biological populations differ. Thermodynamic systems

tend towards entropy while organisms tend towards order (Prigogine 1980, p.83) and possess capabilities that thermodynamic systems do not have. The virtuality of

an organism includes the un-actualized capacity to affect and be affected. Attaching capacities to intensities means that its definition should be extended from a purely

thermodynamic outset. The term intensive should include the processes which give rise to the forming of assemblages. That is, the ability of an object to form relations

with other objects. Extending the concept, a process can be stated to be intensive if it relates difference with difference. Journey

This short venture through the notion of becoming has, as noted above, been made in

an effort to attach a vocabulary to the nature of the uncertain journey between architectural representation and the build; as implied by Evans and Allen. In my reading of

this journey it is necessarily one of becoming. It is a time-based event, an evolutionary

process of diversification and increase in complexity. As already noted, when Allen

places emphasis on the concept of notation, which naturally opposes resemblance, I read it precisely as a technique that addresses the virtual. Thus, notations can be observed as placeholders for virtual components effectuating intensive processes. When the notion of virtuality is attached to the drawing it follows that actuality can be

attached to building. In this the journey from a space of drawing to a space of build-

ing, rather assumes the form of a single space; an environment where individuation through intensive processes and the forming of assemblages define how a build artefact manifests itself. I will propose that, rather than thinking in terms of the three-

part schema drawing–journey–building, then the journey is the main element and the

drawing a necessary device for steering towards the destination of the building, seen as a solution to an underspecified problem. Therefore, the necessary question of representation becomes one of conceptualizing it as a device for steering.

Since this is still somewhat intangible, in the following I will take a route back towards more concrete and direct proposals as to how we might think representation – con-

ceptualized as such a steering device, following a journey of becoming in the form of a design practice.

180


REPRESENTATION AND AGENCY

Reflexivity and Cartography

According to Manuel Delanda the capacity of material and energetic systems to self-

organize and self-assemble is concealed when one focuses on linear causality at the expense of more complex forms. In similarity to how Evans observes the relation

between a conventional practice of drawing and building, classical causality presuppose the characteristic of externality, whereby causes are perceived as external agents

operating on relatively passive targets. These components break down when the object acted upon becomes decisive for the effect.

“(…) while linear causality makes the response of a material system to an external cause basically unproblematic (given the cause, there is nothing else in the

effect that demands explanation), nonlinear and statistical causality re-prob-

lematize material systems, showing them capable of self-organization and self-

assembly, with many things left unexplained in the effect after the mere citation of an external cause.” (Delanda 2005, p.174)

Under non-linear and non-equilibrium conditions matter is intensive and problem-

atic. Drawing on its inherent tendencies, defined by singularities, it is capable of spon-

taneously giving rise to form, as well as possessing a complex capacity to affect and be affected; i.e. reflexivity.

This affects how we should engage with the world. In A Thousand Plateaus Deleuze and

Guattari state that, “it is a question of surrendering to the wood, then following where it

leads by connecting operations to a materiality instead of imposing a form upon matter (…)” (Deleuze and Guattari 1987, p.408). Experimental problems must first be embodied in an intensive assemblage prior to producing a solution, where the practitioner learns

by interacting with – and adjusting to – materials, machines and models (Delanda

2005, p.177). I will suggest that such a performative practice allows us to be directed

towards a mode of thinking with parallels to the cybernetics described in chapter 5. Evidently the non-linear and self-causing vitalism of becoming, described above, has its basis within conventional scientific disciplines such as biology, chemistry and physics – not cybernetics. However, the manner in which Delanda suggests to approach

reality is operationally compatible with the space of the aesthetically potent environment of Pask, and a condition of decentred agency. In support of this view Scott Lash identifies that the vitalism of becoming found in many natural sciences today have directly

181


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

resulted in a renaissance of cybernetics (Lash 2010, p.22).

The significant thing to note is that the approach of both Deleuze and Guattari, Delanda and cybernetics suggests a non-linear and non-hierarchical practice, without an

exteriority from which to impose an input/output equation; consequently, a practice

in need of enabling control, not restrictive. According to Kwinter this approach must assign a heuristic model to practice that is neither exegetic nor deductive. Instead it is

genealogical and cartographic (Kwinter 2001, p.40). It represents a tactical approach that belongs to a transitory and volatile materiality of flow and movement.

“Tactics relies upon and embraces, irruptions, changes – the continually individuating flow of stochastic durations. It is herein that tactics intervenes not

space but in time, and (…) always to produce its adversarial effects within the domain of power, yet without actually opposing or confronting it as such.” (Kwinter 2001, p.123) Cartography

Where does this leave the question of representation? Above I suggested that architectural representations should be envisioned as devices for steering within the space

of the journey between drawing and building. I attached the notion of becoming to

such a space and in my investigation of this concept I reached the proposal of Kwinter that approaching a space of becoming requires a tactical cartographic method of intervention. This last proposal will serve as a stepping stone into the remaining part of the chapter. Here the point of departure is the act of mapping, which will be used to

propose the form assumed by drawing – understood widely – within the framework for design suggested by the thesis.

In a proposal for a revision of mapping practices landscape architect and urban designer James Corner offers a promising insight into what representation could be under

the above circumstances. In his essay ‘The Agency of Mapping’ Corner shows how the

notion of mapping can be envisioned as more than a reproduction and tracing of the given (Corner 1999). Rather, his interest lies in mapping as a productive and liberating

instrument. Mapping possess an agency that “[…] lies in neither reproduction nor impo-

sition, but rather in uncovering realities previously unseen or unimagined” (Corner 1999, p.213). By assigning the notion of agency to mapping Corner naturally defines the map as something pertaining to performance. This performance is necessitated by a

182


REPRESENTATION AND AGENCY

[3]

3. Miscellaneous Mappings. From top left: Guy Debord, ‘Discours sur les Passions de l’amour’ (1957), a map reflecting subjective, street-level desires and perceptions rather than a comprehensive totality of the fabric of Paris. James Corner, ‘Pivot Irrigators I’ (1994), a de-territorialisation of the map by the operations of reframing and cropping. R. Buckminster Fuller and Shoji Sadao, ‘Dymaxion Airocean World Map’ (1954), a variably unfoldable map of the world.

183


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

functional double-sidedness analogous to the one assigned to architectural drawing as noticed by Allen. On the one hand maps represent surfaces that are directly analogous

to actual ground conditions, but they also represent an inevitable abstractness that reveal artificial features that remain unavailable to perception. The analogue/abstract

nature of maps makes them doubly projective, or in the terminology of the thesis, reflexive. Thus, “it both captures the projected elements of the ground and projects back a variety of effects through use” (Corner 199, p.215).

Corner notes that this goes contrary to how mapping has usually been perceived, as a

qualitative and analytical survey of existing conditions that one makes prior to undertaking a new project. Therefore, mapping has typically been observed as an element preceding planning “because it is assumed that the map will objectively identify and make

visible the terms around which a planning project may be rationally developed, valuated and

built” (Corner 199, p.216). For Corner this is regrettable because it makes us tend to

view maps in terms of what they represent rather than how it performs and it makes

us forget to experiment with new and alternative forms of mapping. In fact Corner

believes that: “Mapping is perhaps the most formative and creative act of any design process, first disclosing and then staging the conditions for the emergence of new realities” (Corner 1999, p.216). Thus, maps should be seen as activity rather than completed artifacts.

The agency of mapping constructs a blurring of boundaries between a real and a constructed/participatory world, offering primacy to neither map nor territory.

“Reality […] is not something external and ‘given’ for our apprehension; rather

it is constituted, or ‘formed’, through our participation with things: material objects, images, values, cultural codes, places, cognitive schemata, events and maps.” (Corner 1999, p.23)

Maps are completely mediated products and the nature of their reflexivity is far from neutral or uncomplicated.

I set out to search for a means by which architectural representation could be observed as a device for steering. Certainly, a map, as observed by Corner, is such a device. If

we accept that within an ontology of becoming, practice must be tactical, genealogical and cartographic then this performative concept of mapping points towards a prom-

ising nature for representation in architectural design. Though, understood literally,

184


REPRESENTATION AND AGENCY

maps, even creative and productive ones, depart from observations of a given reality. Consequently, the novelty yielded by mapping extends what is given indirectly by

way of discovery. Design, on the other hand, could be said to seek to do so directly by

producing it within the given. It could be argued that the difference is insignificant, but I believe that the slight dissimilarity in perspective conceptualizes the role of the

designer as more than a cartographer. Therefore, in the following I will return to the writing of Stan Allen who offers a number of ideas related to a productive capacity of

design, which will be used to expand upon the mapping practice suggested by Corner – as a result synthesising a position more appropriate in the context of the thesis. Field conditions

In his essay ‘From Object to Field’ Allen describes a condition within a contemporary design space that he attaches to the concept of field. I believe that this condition can

help qualify and expand upon the mapping practice above by offering an understand-

ing of the movement by which design might project back into the given. The position as well references many of the preceding arguments. It exemplifies a tactical approach

similar to Deleuze and Guattari’s surrendering to the wood, it acknowledges reality as possessing an underlying virtual structure, it occurs without exteriority and relates to acts of steering and decentred agency.

Allen offers the general definition of a field condition as “[…] any formal or spatial matrix capable of unifying diverse elements while respecting the identity of each.” (Allen

1997, p.218). They are loosely bound aggregates characterized by porosity and local

interconnectivity. The internal relationships of parts, determining the behavior of the

field, are more important than overall shape and extent, which become highly fluid. Thus, fields oppose hierarchy, regulating grids, axiality and symmetry, and they are

based on relations, not configurations of form. What is special about field conditions is that they have the capacity to make abstract forces visible. This can be thought of in similarity to how magnetic fields can be made visual by a scattering of metal filings. “Field conditions are bottom-up phenomena, defined not by overarching geometrical schemas but by intricate local connections. Form matters, but not so much the form of things as the form between things.” (Allen 1997, p.218)

When Allen describes the properties and potentials of what he is calling field condi-

185


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

[4]

4. Iron filings around a magnet

186


REPRESENTATION AND AGENCY

tions, it is not intended as a systematic theory of architectural form or composition, but as a catalogue of working strategies, available for appropriation, testing, and adjustment, in practice.

“Field conditions intentionally mix high theory with low practices. The working assumption here is that architectural theory does not arise in a vacuum, but always in a complex dialogue with ongoing practice” (Allen 1997. p.219)

In classical architecture: “Individual elements are maintained in hierarchical order by ex-

tensive geometric relationships in order to preserve unity. Geometry is the invisible scaffold

that at once controls the distribution of parts, but disappears in the building” (Allen 1997, p.219). In contrast, within fields parts are not fragments of wholes, simply parts. Ar-

chitecture as field, is not a closed unity, but a structure that can be added onto without considerable morphological transformations. As such, field conditions are inherently expandable.

In order to understand the relationship between (design) practice and the concept of

field conditions Allen reference so called ‘post-minimal’ artists such as: Bruce Nau-

man, Lynda Benglis, Keith Sonnier, Alan Saret, Eva Hesse, Bary LeVa. The works of these artists shift the perception of the work; ”from discrete object to record of a process of

its making, in the field“(Allen 1997, p.225). According to Allen with these artists, local

relationships are more important than overall form and time and process become decisive variables in creating the work. Their compositional principle is described as “the displacement of control to a series of intricate local rules of combination, or as a ‘sequence of

events’ and not as an overall formal configuration” (Allen 1997 p.225). This is tied to the material used, which consciously makes it impossible for the artists to exercise precise

control. Instead the artists establish the conditions within which the material will be deployed. They give in to the inherent logic of the material itself and let this shape and

calculate its possible configurations, only loosely orchestrated by the artist. We could say that they are aware that practice occurs within an environment displaying its own agency.

Field conditions implies “[…] a shift in emphasis from abstract formal description to close

attention to the operations of making. Questions of meaning and composition are secondary” (Allen 1997, p.227). This connects to certain attitudes towards fabrication and

form-making in contemporary architecture. Allen observes that computer technology

187


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

[5]

[6]

5. Post-minimal art: Barry Le Va, Bunker Coagulation (1995) | 6. Barry Le Va,, Continuous and Related Activities; Discontinued by the Act of Dropping (1967)

188


REPRESENTATION AND AGENCY

makes it possible to imagine a form of construction that advances from part to part, as opposed to one from whole to part. “[…] the joint or the detail is not an occasion to articulate the intersection of two materials […], but is instead the locus of an intensive design energy that proceeds outwards to condition the form of the whole.” (Allen 1997, p.227)

The concept of field should not be understood as a metaphor, but as an actual material

condition, where organization, matter, and making, abandons a conventional opposition between construction and form making. Rather, by searching for an exact and

repeatable connection between the operations of construction and the overall form

produced by the aggregation of parts, it becomes possible to bridge the gap between realization and form-making.

Allen suggests“[…] that architecture could profitably shift its attention from its traditional

top-down forms of control and begin to investigate the possibilities of a more fluid, bottomup approach. Field conditions offer a tentative opening in architecture to address the dynam-

ics of use, behavior of crowds and the complex geometries of masses in motion” (Allen 1997, p.238). The reference here is to phenomena operating at the edge of control such as

swarms. These are field occurrences “[…] defined by precise and simple local conditions, and relatively indifferent to overall form and extent” (Allen p.236).

Representation and the Mechanism of Agency

In a world filled with agency the journey of becoming, between architectural represen-

tation and build artefact, is charged with unpredictability – it is imprecise and does not proceed by resemblance but by difference. Consequently, attempting to produce repre-

sentations that seek to resemble or predict the precise outcome of a design process will fail. The use of an absolute and eidetic form of representation and the assumption of

linearity must give way to a vitalist and non-linear condition of becoming, to which we

require architectural representation to assume the form of a device for steering – that is, both as a vehicle of agency, and an observer of agency. As a collection of ideas, the

above elements suggest how such a device might appear relative to the framework for design proposed through the thesis.

In chapter four I quoted Lars Spuybroek for identifying a convergent phase of selec-

tion and a divergent phase of design as the two main stages by which a generative approach to design unfolds. Something similar is at play here. I see the cartography

of Corner and the field of Allen to describe two components of a single model for architectural representation – respectively addressing the initialization and unfolding

of a design process. The cartography relates to a gathering of both agency and the forms

189


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

to which it is attached, while the field condition supplies a mode by which to think the exercise of agency.

The cartographic approach to representation proposes the concept of the map as a way to

think the medium, and it establishes the design space as a landscape in which something can be discovered, and eventually created. Proceeding cartographically entails that repre-

sentations become collections and selections of agencies – in the form of notation – and

descriptions of forms related to this agency. The appearance of this map assumes a non-

hierarchical horizontality, which by addressing and staging agency allows it to become a tool by which to perform through, as opposed to one of planning and prediction.

The field condition of Allen enters the picture when we seek to activate this performance. The mapping allows us to define and describe local conditions relative to which

we may deploy a process of design. The map represents a field condition. By making this field digital, as opposed to the field of the post-minimal artists above, its agencies may

be simulated. This allows the architectural representation to become more than a map, as well as something other than a description of discrete formal configurations. Instead, architectural representation becomes a record of the processes made possible by the interacting agencies within their environment. In order not to become an absolute repre-

sentation, the activity of mapping should be seen as on-going. This relates to how the representation must be iteratively qualified; firstly, by observing its agency, while making alterations and additions according to its unavoidable shortcomings, secondly, because

of its implied attention to the operations of making, which require empirical testing. Findings in this matter must necessarily feed back into the representation, productively contaminating the digital. As the representation in this way is qualified it evolves as a

device for steering the space between design and making, supplying the designer with an enabling form of control.

In conclusion: I propose to envision what constitutes representation within the framework for design as a device for steering the space between immaterial design develop-

ment and physical construction in the form of a map comprised of concrete form and

abstract notation. The notational aspect allows the map to project agency by way of an intensive operationality from which a reflexive relationship to the build might be established.

Representation thus becomes a technology operating algorithmically, as well as locally defined, and cybernetic.

190


REPRESENTATION AND AGENCY

[7]

7. Programmed drawings by author: A speculative experiment into design processes exploring collections, multiplicity and parallelism. The system producing the drawing is a process in three dimensions. It consists of simple self-organising agents and relations forming between them. The agents move in a matrix of intensities guiding their movement. These intensities can be altered by the user of the system to disturb the movements of the agents, whereby new configurations occur. At any time the user can solidify the traces and relations of the agents into stable configurations.

191


EXPERIMENT 2 | POROUS ASCEND


POROUS ASCEND

The Porous Ascend project investigates how algorithmic and generative techniques allows for the utilization of complex, and by other means inaccessible, ways of devising the schema by which we arrange the parts of an architectural object. It does so through

the construction of a system for designing and physically realizing structures of folded

elements, based on the concept of applying recursion to the geometry of the non-periodic Penrose tiling. Within this process the project also explores questions regarding

the making of bespoke digital design tools, digital production, material behaviour and

assemblage strategies. The project points towards a novel approach to working with, and reproducing, complexity within collections of architectural components. With no

predefined coordinization mapping the ever changing fractal pattern of the project, design and building proceeds by a locally defined generation, identification and pairing of elements. In this way the work demonstrates that we can device structures with-

out reference to a global position, by solely referencing by means of relations between neighbouring parts. This suggests a more locally adaptive and reactive organization of

elements, potentially offering more versatile interfacing capabilities with an outside, and an efficient distribution of specific material behaviour.

The aim of this text is to describe the work, in technical as well as conceptual terms, and to elaborate on the findings and problems resulting from it. Preliminaries

The point of departure for the experiment is the problem of complexity in relation to

the creation of the schema by which we arrange the parts of an architectural object. In order to gain overview, when subjected to complexity, an immediate solution is

to proceed by structuring a design into hierarchical and interlinked subsystems (e.g. Alexander 1964). Though, by thinking hierarchically we are drawn towards position-

ing elements with reference to an extrinsic coordinization; as seen with the use of the modernist grid, as well as the ubiquitous arrays of elements within many projects departing from parametric modelling. Here overview increases, but we might not be able

explore the best possible ways of interfacing and connecting with a world unfolding

in a less orderly fashion at several different scales simultaneously (Gibson 1986, p.9). This is precisely why the possibility to operate by means of more complex principles of organization is an interest of the research.

The described work proposes that within a digital design practice the necessity to fol-

low a regular, extrinsic coordinatization and to separate elements by scale is no longer

193


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

[1]

1. Conceptual model. 3D printed plaster

194


POROUS ASCEND

present. By utilizing algorithmic and generative methods we can establish an approach to positioning and establishing relations between elements within a structure that

is locally determined. That is to say, where elements are not identified and shaped according to a global reference, but are instead associated with locally situated infor-

mation. The reason why this is possible is that algorithmic techniques, by means of

computer programming utilized as a design tool, allows us to externalize the overview of the complexity attached to the formation of a structure from the designer (Terzidis

2006). Hence, relational data are stored and organized within the computer while the designer is allowed to focus on observing the system as a whole.

With this outset the aim of the project was to look for an organizational principle that

could test this idea of a design that continuously unfolds from local points at different scales, and to explore such a principle at the level of the design-tool, production, assembly and material-behaviour. In extension to the previous chapter, the experiment

seeks a field condition, where overall form emerges out of conditions established locally (Allen 1997).

Research/Experiment Context

The final physical demonstrator of this experiment, described below, was commissioned for a public event showcasing research investigating digital modelling and pro-

duction technology within contemporary architectural design. A limited timeframe, economy and the requirement to be able to clearly communicate the project to a broad

audience were decisive for the choice of realizing the system as a tower, as well as the amount of elements, selection of material and included parameters.

The title, Porous Ascend, refers both to the visual appearance of the final demonstrator and the idea that the concept of porosity in a sense addresses the topic of interface and connectivity above. The concept of porosity is a term denoting a measure of the

void spaces in a material. It relates to a notion of relative scale between components of a material. Where porosity is high, the difference between scales of individual com-

ponents is small. Low porosity, on the other hand, requires a significant variation in sizes to fill the space enclosed by the material. In the context of the project porosity figuratively speaking becomes a measure of connectivity.

195


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

System The Pattern

As a starting point for the research, and in order to create the desired situation for

the distribution of elements, an underlying conceptual pattern was invented. This pat-

tern brings together two geometric concepts: recursion and the aperiodic Penrose tiling.

These are of interest because each possesses a property of the sought after principle of the locally determined and a seamlessness of scale, and in combing them a formal situation arises, where the organizational situation of interest can be explored.

In mathematics recursion is the act of defining an object in terms of that object itself (Weisstein 2013), in computer science this becomes a function written in such a way

that it calls itself from within; when the idea is applied to geometry a shape acts as placeholder for the drawing of similar shapes, often as scaled versions of the original

(Flake 1998). Fractal geometry is naturally recursive. This type of geometry represents

a flatness whose singularity lies in its ability to maintain an extraordinary but constant level of complexity at every scale. Sanford Kwinter formulates it thus: “A fractal object is complicated because it contains the principle and rule of its being at every point; it contains

an infinity of stages and cycles, and, instead of occupying new unitary dimensions, it inhabits

the infinitely variable space between dimensions� (Kwinter 2001, pp. 129-130). In a sense we are presented with a seamlessness of scale, which is the particular that is of interest for the research.

Regarding the Penrose tiling (Penrose 1974), the property that is of interest here, is

its inability to exchange one section of the tiling with another. That is, it is an everchanging pattern. This occurs because it is generated through operations of symmetry

rather than translations. The result is a situation where, even though the individual rhombic figures of the tiling repeat, the neighbourhoods, into which they fit, do not.

The idea behind the pattern developed for the project was to apply geometric recur-

sion to this ever-changing pattern of the Penrose tiling, so as to supply its continuous unfold with the ability to act at different scales, in a fashion where shifts in scale occur instantaneously without the use of gradient transitions in the geometry. This was

inspired by previous projects utilizing recursion but which departed from a hexagonal pattern (Aranda Lasch 2008, Tamke Thomsen 2009).

In order to utilize this pattern an appropriate design tool was needed by which a de-

196


POROUS ASCEND

[2]

2. The Penrose Pattern subjected to recursion

197


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

sign process could unfold. The overall criteria for this was to be able to interface with

the generation of the pattern, utilize it as a spatial structure and to unfold this struc-

ture for fabrication. As in the Lamella Flock project I decided to write this tool from scratch. In its outset this allowed the tool to depart from the specific situation of the

project, rather than a general purpose tool-set. This appeared appropriate because the dialog between designer and structure required the development of a customized user

interface. Additionally recursion rapidly challenges computer speed and early testing in Rhinoceros and Generative Components proved very time consuming compared to the results achieved through code written using the Processing language ( Java, http:// processing.org/). The Tool

The bespoke design tool is divided into a viewport, displaying the three-dimensional

design, and a graphical user interface (GUI). Through the GUI the user is given ac-

cess to a series of sliders and buttons that makes it possible to manipulate the design; e.g. amount of recursion, extrusion, curvature and positions of openings. A principal technique for defining variables within the design is here the ability to link 2D grey tone images as placeholders for 2D arrays of information. This means that where the

manipulation of single numerical values is not meaningful, interfacing occurs through images. Here values are extracted by reading the grey-tone-levels of the image at a po-

sition extracted from coordinates within the structure being designed. As an example, at the level of the underlying 2D pattern such an image is used to decide where the pattern performs recursion and for how many levels.

Turning the 2D pattern into a 3D structure can be performed in two different ways: Either as a ‘landscape’ where 3D elements are extruded upwards from the pattern (allowing for overview of the curvature of the overall surface), or as a ‘tower’ where elements are extruded outwards from a cylinder that the pattern has been wrapped

around. The reason for doing so is that the cylinder presents a generic base for the tower shape, that allows a bottom-up behaviour of the elements to define the designs overall shape, while keeping to this category of structure.

The individual 3D elements are shaped by connecting the end and start point of the

extruded edges to the extruded centre point of the rhomb; drawing a triangular surface. This was done for both back and front side.

As with the amount of recursion the shapes of the individual 3D elements were also

198


POROUS ASCEND

[3]

[4]

Recursion (amount and position)

Openings

Topography (inner/outer surface)

Continuity

Component speacing

Infill topography

Centre extrusion (concave/convex) [5]

3. Screenshots of the design tool in use | 4. System parameters | 5. Extrusion of an element

199


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

informed by interfacing with grey tone images. Here the variables controlled were: 1. Extrusion of elements from the 2D pattern: This was done for both front and back of elements relative to the 2D pattern.

2. The amount of sloping as opposed to a stepwise extrusion: This made it possible to create transitions between a fluid and a crystalline formal expression.

3. The position of the centre point described by the triangles forming the front

and back of an element: That is, varying whether the expression of the element was concave or convex.

4. Opening up and closing off of elements: This made it possible to create transitions between elements being frames as well as solids.

5. The amount of offset between elements: Making it possible to have gaps within the structure.

6. Controlling the height of a possible frame within the offset. (The last three variables were not included in the final demonstrator due to time and cost constraints.)

The Making

The aim of creating the above described tool was twofold. Firstly, as a means to test the

overall framework of a design system operating relative to a locally determined struc-

ture, and secondly, also to develop it towards making the designs realizable as physical models demonstrating that the system could in fact relate to the build. For this specific purpose the system was developed in such a way that the geometry of each individual

element can be laid out in two dimensions as an unfolded component. This allowed the structure to be producible through CNC cutting of a plate material. Preparing the data

The most challenging problem in dealing with a variably recursive pattern lies in the

200


POROUS ASCEND

fact that no pre-defined coordinization mapping the position of elements exist. This means that indexes are given relative to where and when they are drawn. The point being, that position is locally, not globally, defined. This means, in effect, that the indexes of neighbours to an element need to be extracted as an additional operation after generating the structure. Getting the indexes of neighbours to an element is the only

way to identify where in the structure an element is situated and therefore impossible to build without.

The technique applied to this problem was to subdivide the edges of each element

according to its level of recursion. This was necessary because large elements would potentially have more neighbours than small ones. For each subdivision an algorithm

was applied to search the population of elements for coinciding edges. When a co-

inciding edge was found, the index of the element it belonged too could be extracted and attached to the edge searching for neighbours. By following this procedure for all

elements, all edges and all subdivisions of edges it was possible to map the connectivity of the entire population of elements.

Unfolding for production and building

The unfolding of the elements forming the structure was also done by means of writing a piece of software. The challenges attached to developing this tool was:

1. To avoid writing a universal ‘unfolder’ by deciding the best strategy for unfolding the specific elements within the project.

2. To deal with the problem of material tolerances and thickness Regarding the strategy for unfolding the individual elements, it was assumed that efficiency would be compromised if unfolding was done in the most appropriate way

relative to each individual element. This would have meant that the geometric configuration of each unfolded element would not always be the same. In this case the folding process – back into physical three dimensions – would have been slowed down

by requiring an interpretation of the unfolding strategy creating the piece. Instead

a strategy was devised where approximately 95% of the elements could be unfolded

201


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

[6]

[7]

6. Process | 7. Line drawing of the final demonstrator | 8. Unfolded elements

202


POROUS ASCEND

[8]

203


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

in the same way1. The remaining 5% required a slight manual redraw to avoid self-

intersecting geometry.

Material tolerances and thickness were also a challenge. In order to accommodate for imprecisions occurring from both folding the material and the thickness of the

material itself all unfolded elements where reduced in size according to measurements gained from prototyping in 1:1

Additionally it was necessary to attach information visually to the sides of all elements regarding neighbours. This meant the index of the correct neighbour, as well as the

contour of the side of this neighbour. This last feature was necessary because elements would often be extruded in steps relative to each other, i.e. neighbouring elements were often not connected at the corners, but somewhere along the surface of a side. Production

The unfolded elements were collected in a CAD file and nested to fit cardboard sheets of 1 x 2 meters. This sheet size was decisive for the maximum size of an element, since it was unfavourable to divide an unfolded element into two pieces.

Different colours were used to divide lines into cutting, folding and drawing paths. The last type was needed for getting information regarding indexes and the exact position of neighbours, mentioned above, onto the cardboard.

The cutting was done using a CNC knife-cutter producing 527 individual elements. Building

The actual building of the structure proceeded by first folding all elements and sorting them in numerical groups for better overview. A 2D map representing the position

and numbering of elements was produced to keep track of the assembly process. From

here sections of the tower were identified and assembled into larger fragments, these 1 The strategy was as follows: The first triangle making up the top of an element gets re-drawn in 2D by measuring lengths of, and angles between, segments. Subsequent triangles forming the top gets redrawn in 2D in the same way, and rotated so as to connect to the previously drawn triangle in the order that they are drawn in 3D. From each of the triangles forming the top part of an element its corresponding side is drawn, also by using lengths and angles measured in 3D. In connection to the first of these sides the bottom part of a triangle is placed, drawn in the same way as the top piece, but with the first triangle rotated so as to fit the end of the side-polygon. Should it turn out that the unfolded sides are intersecting, this strategy is repeated starting from the next triangle forming the top part of the element.

204


POROUS ASCEND

[9]

9.-12. Final demonstrator

205

[10]


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

were again brought together to form the actual tower. An interesting material prop-

erty, resulting directly from the use of recursion, was observed in this process. In areas

with high amounts of recursion it could be detected that structural strength followed. In this way the utilization of recursion as a technique for the arrangement of architectural components points towards a potential for a distribution of structural strength relative to locally defined needs.

The final structure stood approximately 4.8 meters tall, when all 527 elements had been folded and connected according to the production drawings.

Findings and Challenges

The experiment exemplifies how the use of an algorithmic and generative logic al-

lows for an understanding of component-based structures, which does not require the

determinism of a globally determined system of distribution. Another way of think-

ing is produced – and required – by which design operates through mechanisms of

local connectivity and relation. If we perceive the concept of structure as a threshold

between an exterior environment and interior organization – as within the autopoietic system described in chapter 5 – I will suggest that this way of arranging a structure

holds the potential to effectuate a cultural shift in how architecture might relate to its environment. Rather than operating by the regularity of the sequential, we are allowed to proceed, not by origin in scale or position, but relative to the qualities of

densification and continuous unfolding. This implies a possibility of immediate transi-

tions within a structure either reflecting external conditions or an internal projection towards the surroundings. In this way the contribution made by the experiment – to the framework for design proposed by the thesis – is that by utilizing algorithmic techniques we may exchange an absolute schema with a generative, as the means by

which to relate, arrange and distribute the parts of an architectural object; thus, in-

creasing variety, adaptability and reactivity of a structure, by allowing access to a higher degree of complexity.

Beyond its formal and diagrammatic potential Porous Ascend also introduced new

productive problems into the framework of the thesis. These come to light by way of

the shift made within the experiment from digital system to physical demonstrator. Through the process of erecting the physical structure it became clear that its material basis does not act as a passive receiver of form. The demonstrator problematized the

206


POROUS ASCEND

pattern and the derived structure by exhibiting how they belong to a somewhat closed

geometric/ideal space, untouched by the reality of the material. Thus, while the project can be used to hypothesise an improved capability to interface with an environment, the concrete environment of the material manifestation of the structure should preferably have been implemented to a higher extend within the design space. In the

demonstrator the geometry of the system meets something that cannot be restricted

to that geometry alone. It encounters something other than itself, by which its complexity increases, but which is separate from the design space. Consequently, a practical and potentially productive property is placed beyond the control of the designer.

Even though digital modelling allows a significant precision within the digital design space, this precision is always met with the necessary imprecision and unruliness of a material reality. Material thickness, properties and handling all add up to produce a situation where we can never assume the kind of exactness offered by the purely

digital. Thus, what is required here is a phase of empirical research to which to relate

the overall structure, as a means to extend the journey from digital representation to physical demonstrator.

[11]

207


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

[12]

208


POROUS ASCEND

209


7 | ENVIRONMENT AND INTERFACE


ENVIRONMENT AND INTERFACE

In chapter six I proposed to understand that which constitutes representation – in a

contemporary and generative digital practice – as a cartographic device for steering, which operates relative to a notion of agency. This was necessitated by the fact that the contemporary digital design spaces observed by the thesis do not operate relative to

stable and absolute models of form, process and context. Rather, these spaces challenge a notion of absolute representation, suggest new degrees of complexity and make new

processes possible. Non-standard critical practices of a digitally produced architecture operate in a field of topological diagrams, interdisciplinary information flow, nota-

tional systems and material performance orchestrated by computational, algorithmic and generative technologies. In addition to the consequences that this produces with

regards to questions of representation, I believe that this also implies a need to critically consider the role played by the concept of context in architectural design.

Any design space must necessarily operate relative to a larger contextual space to which

it contributes and relates. In a sense this space is simultaneously a component of the

design space and external to it. The cartography suggested in the previous chapter, and

the field it is targeting, of course implies the presence of such a double projection rela-

tive to a space of exteriority. Hence the practice of mapping, and the particular agency

attached to this activity, designates a principle exemplifying the conflicting duality

of a field that is at once internal and external to a process – as well as a product – of design. The aim of this chapter is to address and expand upon this space, through

which a practice might steer. I will make two proposals in this endeavour. Firstly, that

the underlying field of a design space becomes conceptualized relative to a notion of environment and that this conception in fact dissolves the divide between context and

object – merging the space of exteriority, medium and form. Secondly, that this un-

derstanding of a design space becomes operational through the notion of interface – a principle I define as the ability and function of components to communicate while maintaining a degree of individuality.

Environment without Exteriority

Throughout the chapters of the thesis I have presented a number of positions that I

believe implicitly challenges an architectural concept of context: Stan Allen urges us to

operate within ‘the field’; Robin Evans directs our attention towards a space between representation and material reality; Gordon Pask operates performatively from within an aesthetically potent environment; Scott Lash describes an algorithmic and completely

211


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

mediated reality. I believe that when one observes the overall body of these ideas a design space conceptually appears to assume the guise of, what might rather be called

a design ecology, in reference to natural systems of interrelated becomings. In such a perspective of constructing a surface onto which a design might develop, instinctively

the use of a concept like environment seems more appropriate than that of context. However, in these matters intuition will not suffice. I will therefore begin this section by shortly examining the notion of context to be able to situate the present discussion

in relation to this concept. I do this in order to argue that this concept possesses a binary relationship to architecture, which I will propose is undesirable within the inquiry of the thesis. Context

In his book ‘Words and Buildings’ professor of architectural history Adrian Forty traces the main development of the concept of context in architecture (Forty 2000). I will

depart from his account in the following. The word ‘context’ was introduced into the

architectural vocabulary in the 1960s but seems to have evolved out of a critique of modernism since at least the 1950s. Here the Italian architect Ernesto Rogers coined a

term with the name le preesistenze ambientali. In English: the surrounding pre-existences.

It was used as a tool in a first wave of objections to modernism. A movement that, as

observed by Forty, was criticized for treating all problems as unique abstract, possess-

ing an indifference to location, and in the need to always make a prodigy out of works of architecture (Forty 2000, p.132). Rogers argued that architecture should be in a dialogue with its surroundings, both in a physical sense and in awareness of historical

continuity. Important to the term of Rogers was the historical continuity of the city. Here formalism is believed to take hold when one does not absorb into a work the

particular and characteristic contents suggested by l’ambiente. To consider l’ambiente

means to dialectically take history into account. Moving ahead in time, Forty notes that Aldo Rossi criticizes Rogers ambiente for being un-concrete, and instead proposes

to make it thus by studying architectural forms themselves, independently of their functions. This, because he believed that a contact between the processes from which the city emerges, within the cultural and historical, directly exists in form.

Le preesistenze ambientali was wrongly translated into English as ‘context’ but that word does not actually appear until 1964 in Christopher Alexander’s “Notes on the Synthesis

of Form”. Here context is used synonymous with how Forty defines the concept of environment; as something that describes the surrounding objects. For Alexander form

212


ENVIRONMENT AND INTERFACE

is the solution to the problem and the context is what defines this problem; this analytical approach to context has been discussed earlier in the thesis. Later on the term

is used by Colin Rowe who concentrated on the more formal properties of works of architecture and was, in contrast to Rogers – who perceived the context as something

formed by objects – interested in the relationships between objects and the space they occupied. Rowe proposes a so-called contextualism; a practice of explaining how idealized forms can be adjusted to fit a context. This was part of a project of creating a compromise between the modernist and the pre-modernist city.

From the above we can easily conclude that context is not a clearly defined term. Or at least, different interpretations can be observed, where its specific constituents are chosen relative to personal ideologies. Wherein lays the commonalities, if any, then?

Whether context consists of objects, aesthetics, relations, or a historical positioning it seems to always address something that surrounds the build. What we see, is a binary

relationship that, according to individual positions, is more or less dynamic. The surroundings, as described in the theories of context above, are always possible to think

independent from architecture; as a pre-existence. It is in the above always a question

of a model possessing a duality. We can observe an opposition to such a binary relationship in the position of Deleuze and Guattari, whom by departing from readings

of biology and zoology find a clear division between form and surroundings incon-

ceivable. Here we find a mode of thinking where boundaries between environment

and form, and forms and forms, are not necessarily clearly traceable (Deleuze Guattari

1987). In this Deleuze and Guattari represents an ontology within which context

can be thought differently from the above and I will suggest that this is a difference that might help us develop a more meaningful relationship between a morphogenetic process and its surroundings. I will therefore return to the position of Deleuze and Guattari below.

The current investigation, should however not be read as an attempt to eliminate the concept of context, rather it is a reflection on how it can be developed and how it

can be transformed to accommodate the specific focus of the present research. In recognizing that the research pursues a possibility of a loss of hierarchy by means of mechanisms of reflexivity, using a term that is culturally biased to point towards a duality does not seem to further the investigations. By introducing environment

instead, it is believed that we conceptually deal with a term that cannot be thought of independently from the forms it surrounds.

213


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

Environment

Regrettably the concept of environment appears to possess a vagueness comparable to that of context. It is employed within diverse fields of inquiry, which all seem to assume an unspoken consensus concerning its definition, or maybe rather, a demarca-

tion of its use. From one position the nature of an environment could be given as; the surroundings of those organisms that perceive and behave (Gibson 1986), in other cir-

cumstances the environment acts more as reservoir of energy exchange (Von Foerster 2003b). Consulting a dictionary offers a definition of environment as something along the lines of the aggregate of surrounding things, conditions, or influences. But this easily assumes the environment as pre-existing its constituents. Within architecture ‘environment’ might for some solely point towards environmental issues, because context

already acts as the placeholder for aspects of the surroundings. Neither of these uses of the word environment is what I mean to suggest here. Instead I posit that the concept

of environment designates a dynamic space that simultaneously forms and is formed by a process of design. A kind of folding of the surroundings, that makes them imma-

nent to the design space and what is designed. Environment thus loses its exteriority. In chapter six we saw something akin to this in the doubly projective mapping practice of Corner, where representations always reflect back on what is represented.

This resonates with certain tendencies within digital design practice. Mark Goulthorpe observes that form-generating processes have the ability to relate to an environment. He defines two modes that such ability can proceed by; alloplastic or auto-

plasic. Autoplastic is defined as a self-determinate operative strategy, while alloplastic is described as a reciprocal environmental modification. Goulthorpe hypothesises that digital architecture operates within a now alloplastic space, as one begins to work in a

responsive, conditional environment, sampling and editing the proliferating capacity of

generative processes, which offer a transformative creative medium by its very nature (Goulthorpe 2008). In a manner of speaking generative processes make little sense

without the presence of an environment. The process of self-organization, for example, is according to Heinz Von Forester inconceivable without the presence of – and relation to – an environment:

“I propose to continue the use of the term ‘self-organizing system’, whilst being

aware of the fact that this term becomes meaningless, unless the system is in close

contact with an environment, which possesses available energy and order, and

214


ENVIRONMENT AND INTERFACE

with which our system is in a state of perpetual interaction, such that it somehow

manages to “live” on the expenses of this environment.” (Von Forester 2003a, p.3)

We need to understand that form-generating processes are conditioned by a depen-

dency on an environment. In relation to the field of architecture a concrete proposal as to the nature of such a dependency and an exemplification of environment as something different from context can be illustrated by the essay ‘On Correalism and Biotech-

nique’ written in 1939 by Frederick J. Kiesler (Kiesler 2007). The text departs from the hypothesis that functionality evolves through the interaction between man and envi-

ronment, but not as a direct result of environment. It is rather the development by the

environment of something which was already inherent structurally in the architecture. Consequently, all functions and all forms are contained in structure. Kiesler therefor proposes to replace the concept of ‘form follows function’ by the progression of structure, function, and form. Kiesler’s view presents an interesting similarity to the theory

of autopoiesis, where structure is precisely something that mediates between environment and program (organisation), and it is the interpretation of the result of this

mediation, as observed within an environment, that gives rise to a notion of function. Thus, in both the case of Kiesler, as well as Maturana and Varela, structure is a priori

to function and function is always dependant on observations within an environment. Kiesler relates architecture to a network of forces, where form is only the visible representation of integrating and disintegrating forces. Reality consists of these two categories of influences which are in constant interaction. This exchange of interacting

forces Kiesler terms ‘co-reality’ and the science of its relationships, ‘co-realism’. This term

expresses a continual dynamic interaction between man and environment. From a perspective of human activity Kiesler believes that environment cannot be defined in

biological, geographical, or socio-economical terms alone, wherefore he proposes the additional notion of a technological environment.

Kiesler perceives technology as an evolving entity governed by a ‘need-morphology’. He

believes that industry is not that which is producing the technological environment but that the technological environment is produced by human needs. The technologi-

cal environment consists of systems of tools, defined as any implement created by man for increased control of nature. Every technological device is co-real and hence conditioned by its relation to its total environment.

215


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

Design based on form following function is according to Kiesler flawed because the functional object is not driven by environmental stimuli arising from the actualization

of needs. Instead he proposes a mode of design termed biotechnique whose object is reactive, that calls for the need to understand the methods by which nature builds.

In the present endeavour Kiesler exemplifies an architect with a practice of building

and a need for a concept of environment. In this role I utilize Kiesler as a stepping stone in order to propose my own model in this matter. Kiesler suggests a number of

interesting ideas. Firstly, the assumption that form-generating processes are conditioned by a dependency on the environment in which they take place; secondly, that

both this environment and the structure of an architectural artefact exhibit agencies which displaces a locus of control and the origin of form; and thirdly, that technology

and its development is itself related to an environment. Technology forms systems of

tools and techniques, which are conditioned by their association to an environment. Thus, to form an understanding of the concept of environment also reflects back on our

understanding of technology. In this way the relevance in contemplating the concept of environment lies not just in an ability to better understand and formulate the pro-

cesses by which to develop architectural form, but also in the possibility to formulate principles for evaluating and developing technology.

Environment as Double Articulation, Algorithmic Structure and Virtual Being

In the following I aim to present a conceptual model by which to think a design space

relative to the notion of environment. That is, to supply a discussion capable of making

architectural practice develop through a specific take on this concept. The intention is therefore not to supply a general understanding or model of reality but to supply a toolset by which a surrounding externality may be mediated and represented by the

digital, and by which this field can become immanent to the design space itself. This

conceptual model will be described in relation to three interrelated constituents – a general form, which I will describe as a double articulation; a structure related to what can be defined as an algorithmic mode of operation; and an overall being attached to the idea of the virtual.

What I aim towards is a situation where the concept of environment may exchange that of the object, in the sense that the expression of form becomes a feature of, and

216


ENVIRONMENT AND INTERFACE

within, an environment. I wish for a perception of the design space in which a binary

model of separation between the architectural artefact and its associated environment

disappears. The account should be understood as a diagrammatic and operational comprehension of the design space. In this endeavour I borrow concepts from outside

the perspective of the present discussion. These do not formally represent ideas about

what they point towards here, but suggest a conceptual potential I use to advance the practice of the thesis.

Double Articulation

It is the notion of double articulation that is the main point of interest in the present

endeavour. This concept, borrowed from Deleuze and Guattari’s ‘A Thousand Plateaus’

(Deleuze Guattari 1987), refers to a diagrammatic operation by which relatively stable systems come to exist and persist. This process follows, as the name implies, two ar-

ticulations related respectively to content and expression. The first articulation relates to

a process of selecting – or territorialising – content from the substance of a flow in a

milieu. This is then subjected to acts of coding, which represents the form given to the content. The second articulation relates to a process of folding, by which a form and a

substance of expression result in new relations and entities with emergent properties

(Bonta Protevi 2004). This is highly reminiscent of the way James Corner describes the

agency of mapping, as a device projecting back into what it represents. Thus, within the concept of double articulation lies a potential to expand upon, and relate to, the notion

of reflexivity. Double articulation allows for an ability to form emergent unities which

nevertheless respect the heterogeneity of their component parts, which relates to how I understand the concept of interface. Deleuze and Guattari offer the following description of this idea, borrowed from nature:

“The orchid de-territorializes by forming an image, a tracing of a wasp; but the wasp re-territorializes on that image. The wasp is nevertheless de-territorialized, becoming a piece of the orchid’s reproductive apparatus. But it re-territo-

rializes the orchid by transporting its pollen. Wasp and orchid, as heterogeneous elements, form a rhizome.” (Deleuze Guattari 1987, p.10)

Many elements within this quote require further explanation. However, before entering into more detailed elaborations I need to explain its relevance. This can be done in

the form of a question: What if a similar model were to describe the relation between archi-

217


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

[1]

1. Ophrys Apifera, Orchid

218


ENVIRONMENT AND INTERFACE

tectural medium and architectural artefact? A great deal of agency exists within the image

conveyed by Deleuze and Guattari. If we were to think of the agencies of the digital medium, as a mapping in the sense described in chapter six, and its derived material or

visual evidence in terms of a relation of double articulation, I will suggest that we are

supplied with a device for counteracting a reversion to the binary model I attach to the concept of context. That is, by utilizing the concept of double articulation we are able

to identify when a digital practice repeats a binary relation by developing models that themselves end up as contexts in the sense described above. That is, where mapping

and planning result in inanimate pre-existences within the design space, which are not affected or evolve together with the forms of expression developing within it. Virtual

The double articulation is the significant event in that it is the operation that permits the actualization of material form, from within an environment. However, in order

to understand this process of actualization – and to set it in motion – a description of an underlying un-actualized state of being is required. I referred to this above as the flow of substance, which the first articulation of content engages. Concretely,

what I aim to describe is a notion of the virtual, since it is the virtual content that

is territorialised/coded when a form is expressed. The relevance of defining a virtual state of being is that it can be used to propose a condition within a design space, in which the appropriate structure and model for a reflexive double articulation be-

tween a design environment and a design outcome can be played out. Importantly,

the virtual should not be considered as constituting the environment here. Rather, it should be understood as the support for a structure through which information can

flow. I observe the associated environment of a design process as something depart-

ing from the territorialisation of content within this flow. The point here is to suggest that practice within a digital design space, perceived as an environment, departs from a non-essentialist un-actualized state of being, which can be described as a complex structure that does not reference to a priori unity.

In his book ‘Genesis’ Michel Serres observes a fundamental problem that compli-

cates an endeavour to think the unformed (Serres 1995). Namely, a strong cultural urge towards unity characterized by a need to classify, codify, and solidify concepts

and forms. The consequence hereof is that the possibility of change becomes lost. By going beyond the need for unification, which is necessarily already codified and ex-

219


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

pressed, Serres encourages us to think relative to its opposite – multiplicity. The mul-

tiple should be understood as a hybrid state of being between blank absence and chaotic abundance (Serres 1995, p.63), or as a field of un-directed flow. Serres proposes

to perceive the world, and becoming, as movements relative to the multiple, to struc-

turally think reality as originating from the disordered state of being in which we find a multiplicity in its purest form. From this condition the multiple can be thought of,

as and in itself, without the aid of the solidifying concept of a unity. However, Serres is capable of relating the multiple to the unitary by describing the multiple through

the overall diagrammatic representation of a viscosity. The idea behind this diagram

is to think the becoming of form as a process of solidification, where time should not be measured in extend and duration but in varieties of paths to follow. Thus, in order

to supply a space for a becoming of form – i.e. an environment – a capacity to possess indeterminacy that later in a process gives way to a condition of singularity overflow is required.

Thus, I will propose that a successful design environment depends on an ability to

establish a point of departure characterised by a productive indeterminacy capable of diversifying the range of possibilities a process of design might pursue, as it propagates from mapping to architectural artefact. Abstract Machines

While a state of indeterminacy suggest the appearance of a productive variety within the substance of a design environment, a more precise notion of structure and instru-

mentality is needed in order to establish the proposed reflexive double articulation

between medium and artefact. From the un-directed complex structure of the virtual another needs to form from which the selection of content may lead to expression. To

this end Deleuze and Guattari describe acts of capture within the vitual which they call stratification (Deleuze and Guattari 1987, p.40). These are topological territorializing operations. That is, a coding or ordering of a substance. A surface of stratification

acts as an assemblage between layers of substance that appears as a slowing down, or

thickening, of the virtual field. The pure substance of the virtual lies outside the strata, which tends towards actuality. This pairing of layers exhibits phenomena constitutive of double articulation. Firstly an articulation chooses or deducts from a flow of

substance upon which it imposes an order, connections, and succession. This is giving

form. Secondly an articulation establishes functional, compact, stable structures (i.e.

220


ENVIRONMENT AND INTERFACE

[2]

2. Fluid simulation (Stam 2003): Experimental probe by author (a very direct examination of the concept of viscosity). This fluid solver is based on the so-called ‘Navier-Stokes’ equations, which are precise descriptions of how most fluids flow in nature. Mathematically, the state of a fluid at a given instant of time can be modelled as a velocity vector field. That is, assigning a velocity vector to a grid of points in space. This vector field is constantly changing due to heat and pressure and its own behaviour. To visualize fluid we need to visualize something akin to smoke or dust particles. Since it is beyond the capabilities of computers to do real-time simulation of millions of particles, as they appear in reality, we can, though, visualize the densities of these. What is done is to parameterize a portion of space and calculate the density in it at a given moment. The solver consists of a number of functions that moves densities and velocities. Moving the densities consists of following the vector field, diffusion, and operating a source producing the densities. Diffusion occurs as a spread of densities between grid cells. The cell’s density decreases by losing density to its neighbours, as well as increase when densities flow in from them.

221


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

forms) that lead to the development of the substances needed to actualize these forms. As mentioned above the first articulation concerns content, the second expression, the actual distinction is therefore not between form and substance but these two.

A question that imposes itself is how the double articulation of content and expres-

sion passes from a virtual to an actual solidified state of being. Deleuze and Guattari propose that things actualize through the ordering principle they call ‘abstract machines’. These entities are vital participants in the event of double articulation. I will

also suggest that they are reminiscent of the concept of algorithms – that is, rule-based procedures dependant on the environment in which they unfold.

”We define the abstract machine as the aspect or moment at which nothing but functions and matters remain. A diagram has neither substance nor form, neither content nor expression.” (Deleuze and Guattari 1987, p.141)

There are rules, planning and diagramming, attached to the abstract machine, not

randomness. They are machines that constitute becomings. In the example of De-

landa, discussed earlier in the thesis, abstract machines would be termed singularities, when appearing in a context of natural sciences; that is to say, entities defining the

long-time behaviour of processes in complex dynamical systems. Thus, virtual multiplicities serve as structures for such processes. By making the abstract machine part

of the virtual dimension of reality, means that this state of being is neither an undif-

ferentiated aggregate of unformed matter, nor a chaos of all kinds of formed matters.

Abstract machines shift between the two states of territorialisation and de-territo-

rialisation. That is, they exist either as relative to the solidified manifestations of the virtual – that is the strata – or abstractly within the virtual itself. The latter being their

primary condition, thus, making the virtual always immanent to actual forms. The distinction between the territorialised and de-territorialised state of the abstract machine is related to what is constant and what is variable in a stratum from the perspective of

substances and forms. Again, using the image of the algorithm this is comparable to the difference between an algorithm ‘as such’ and a single instance of it in use.

In contrast to strata (actualized structures/components), within the virtual dimension there are no forms or substances, content or expression, or relative movements of de-

territorialisation. Here the role of the abstract machine is to construct continuums of

intensity; it creates continuity for intensities that it extracts from distinct forms and

222


ENVIRONMENT AND INTERFACE

substances.

Machine assemblage is the operations performed by the abstract machine on a stratum. Here it performs co-adoptions of content and expression and ensures the relation

to whatever serves as a substratum and brings about the corresponding changes in organization. Lastly it connects with the virtual dimension because it effectuates the

abstract machine on a particular stratum, between starta, and in the relation between the strata and the virtual. This indicates a continuous negotiation between the virtual and the actual, which in effect suggests that pure virtuality or actuality does not exist.

Using elements from the above I can expand upon my understanding of the concept of environment and attach to it the notion of abstract machines. I will propose that the

concept of environment can be described as a field, which exists as a gradient between the virtual and the actual, consisting of the elements of form, substance, content or expression – thus being the potential stage of an orchestration of these elements. This

field can be characterised by a productive indeterminacy, which I will suggest must

necessarily be approached by means that possess adaptability as opposed to predictability. As stated above there is something reminiscent of the concept of algorithms in the idea of the abstract machine. Similar to this concept the abstract machine does

not require content or expression, but adapts and develops relative to its concrete posi-

tion within a system. It points towards the singular as well as the multiple. What can

be suggested by contemplating the concept of abstract machines is that the structural entities assisting in the process of digital design within an environment – that which

sets the elements of the environment in motion – are principles of interface, adaption and movement. In a sense this simply restates positions established in previous

chapters. However, a consistency between these positions and the concept of environ-

ment manifests. Thus, allografic projection, notation and decentred agency are means by which to set the abstract machines within an environment in motion. In fact we might think of theses as the design of such machines themselves operating relative to

ones already present in the environment with which we wish to engage architecturally. Open and Systematic Montage

In the above quote concerning the wasp and the orchid the interface between these two entities takes on the form of what Deleuze and Guattari call a rhizome. This is the

form assumed by a network of abstract machines, thus the concept of rhizome allows

223


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

for a conceptualisation of the nature of structure in the present discussion. Originally the rhizome is a decentralized type of root system borrowed from botany. A robust system which functioning reminds us of how self-organizing systems, e.g. flocks, are

able to function even if divided, or when an element fails in the task it is performing. The main feature of the rhizome is that any point within it can be connected to any other point. This eventually means that the rhizome is a collection of connections. All

multiplicities are in this way non-hierarchical systems contrasted with the arborescent (treelike) system, which is a hierarchical structure where: “an element only receives in-

formation from a higher unit, and only receives a subjective affection along pre-established paths” (Deleuze Guattari 1987, p.16). The rhizomatic structure allows us to connect different regimes and different states. The structure of the Rhizome presents several

operations that might prove useful in the present setting, and again points towards the mapping described in chapter six:

“The Rhizome operates by variation, expansion, conquest, capture, offshoots. (…)

the rhizome pertains to a map that must be produced, constructed, a map that is always detachable, connectable, reversible, modifiable, and has multiple entryways and exits and its own line of flight.”(Deleuze Guattari 1987, p.21)

In the present endeavour of thinking the concept of environment the rhizome is note-

worthy as a structure having no object. This characteristic results from its contrast to the arborescent – hierarchical – system, which would naturally reference a point of origin. Thus, to apply an understanding of the environment as a rhizomatic structure or mapping allows for the abandoning of the binary component of the object. James

Corner describes such a mapping as “…a form of systematic montage, where multiple and

independent layers are incorporated as synthetic composite” (Corner 199, p.245). This construction is simultaneously inclusive and structuring, allowing a variety of aspects to be mapped onto the same model, creating a systemic field of interrelationships with the

characteristic of being suggestive of new creative potentials. The strength of mapping as a rhizome is that it does “not ‘represent’ any one thing at all; rather, it might simply array

a complex combination of things that provides a framework for many different uses, readings,

projections and effects […] without beginning, end, limit or single meaning” (Corner 1999, p.246). This openness possesses a potential of using the present conceptualization of environment as a component in an effort to engage in the practice of design without

224


ENVIRONMENT AND INTERFACE

resorting to inhibiting planning: “Instead of designing relatively closed systems of order, rhizomatic mappings pro-

vide an infinite series of connections, switches, relays and circuits for activating matter and information. Hence mapping, as an open and inclusive process of

disclosure and enablement, comes to replace the reduction of planning.” (Corner 1999, p.250)

Environment and the Becoming of Form

In the above I have addressed a diagrammatic and structural conceptualisation which I seek to relate to the notion of environment. In the following I would like to address

the notion of form and its relation to an environment as well, which will also allow me

to return to the notion double articulation that was the point of the departure for this discussion concerning a conceptual model of an environment.

In their perception of form Deleuze and Guattari are obviously not essentialists. This

means that form follows from its association to a population, to which it is affiliated. Forms are in this way more like statistical results, where degrees are no longer mea-

sured in terms of increased perfection, or a differentiation and increase in the complexity of the parts.

“The more a population assumes divergent forms, the more its multiplicity di-

vides into multiplicities of different nature, the more its elements form distinct

compounds or matters – the more efficiently it distributes itself in the milieu, or divides up the milieu.” (Deleuze Guattari 1987, p.48)

In such populational distribution and sectioning of reality we encounter the concept

of an associated milieu. A concept that breaks with the idea of an environment as some-

thing purely external and a good example of the double articulation described above.

Deleuze and Guattari find their inspiration in the work of biologist Jakob von Uexküll, who speculated that the space of an organism can be understood as a reflexively formed

and immanent condition of subjective observation and experience. He believed that objects always remain products of the subject, without an accessible independent existence (Von Uexküll 2009). For Deleuze and Guattari an organic form is not a structure

in itself, but the constitution of an annexed or associated milieu. The milieu does not

225


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

determine the form, but this does not mean that the relation between form and milieu is un-decisive. Since a form depends on an autonomous code, it can only be constituted

in an associated milieu that possesses the complex characteristics required by the code. Here we see an example of a becoming of expression given the presence of the correct

content. Consequently, the associated milieu is always related to a milieu of exteriority, from which it samples content, but a distinction between interior and exterior is no longer perceptible or meaningful from within it.

The Design space as Environment

The aim of this chapter has been to discuss the space forming between the external and internal relations defining a design space. The proposal in this endeavour is to suggest

that the underlying field of a design space becomes conceptualised as an environment, defined as a space in which the divide between object and surroundings dissolve. Thus ‘environment’ is not a concept of exteriority but a dynamic space, both formed by and

forming a process of design. It represents a folding of an externality which makes it immanent to both the design space and its forms. This produces a certain agency within the design space which displaces control and origin of form.

In order to supply a conceptual toolset by which an externality in this way can become mediated and represented digitally – as something immanent to a design space – I

propose to operate relative to three overall constituents; double articulation, virtuality and algorithmic diagramming. This is hypothesised to create a condition beyond bi-

nary relations, of productive indeterminacy and adaptive as well as reactive structuring, without prediction and object.

I began this chapter by problematizing the concept of context as being representative of a binary relation between a surrounding pre-existence and the architectural artefact. I will propose that by conceptualising the design space relative to a notion of a responsive and conditional environment we are presented with an understanding and structure, where a mechanism of reflexive coupling makes a clear distinction between

surroundings and object difficult. In a sense the object is replaced by the environment, as a moment of solidification within a flow of information. However, the key ele-

ments of this model – externality, internality and form – are complexly linked because

the presented concept of environment assumes a kind of double role. Hence, when I introduce this concept, in relation to the aim of the thesis, it is to be understood as

226


ENVIRONMENT AND INTERFACE

something mediated and represented through the digital design space, as well as an

operative model of this space itself – an understanding of both external and internal structure. I will suggest that the design space as a medium territorialises an environment of externality, thus forming an associated environment. This is consistent with

the contemporary informational models discussed in chapter four, where the agency

of the architectural model takes form amidst relevant informational patterns and flow. I will also suggest that through the containment of its associated environment the system of the design space may become operational as an environment itself. Thus, I

seek to address and cancel out two binary pairs; that of context and design, and that of architectural medium and form.

The concept of environment both contains, and becomes contained, by the design

space. Here the design space is at first a feature of an environment, but by establishing

a projection of this environment within itself the design space creates the condition for producing a difference within its external environment – through the appearance

of new forms. This model necessitates a certain understanding of architectural practice, in that a process of design will need to be observant of this double movement. I will hypothesise that a digital practice is especially well suited to address the complexity

of this situation. Hence, diagrammatic features and operative mechanisms, similar to those discussed above, can be utilized to supply an understanding of how a digital

and generative design space may be developed, but they also supply an understanding

of how this design space is coupled to a material, cultural and technological outside. In this the concept of environment prescribes a specific kind of design system. This

system appears to synthesize many of the elements introduced throughout the thesis. It represents a qualitative perception of technology as a locally determined agency

instigating movements between the intensive and the extensive, it requires topological

diagramming and informational flow, and it decentralises control. Also, it represents something relatable to the field of Stan Allen and a doubly projective mechanism

reminiscent of James Corners cartography. These references reminds us that if we were

to delimit the application of the concept of environment to only denote the internal

structure of a design space, it would conceal the reflexive potential inherent to this conceptualisation. Here a notion of environment would become merely imagery. That

is, the concept of environment can easily be used to produce a design space in the form of a generative ecology which becomes nothing but a formalism replicating the binary

relations I associate with the concept of context. In order to counteract this we must

227


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

continuously maintain interface between the construct of a digital design space and the material, cultural and technological externality to which it is reflexively coupled.

This directs attention to how we should perceive the computer as a medium and the flow of information representing the emergent architectural object. Here I view the

model of the wasp and the orchid, presented earlier, as a means to think the relation between these two entities. If we were to assume a similar reflexive relation between the digital media of an architectural practice and the material forms it aims to describe, then this will have necessary implications for a perception of the computer as

a design tool. Most importantly, that the medium of a digital design practice must

not end up becoming a context in the sense described in the opening of this chapter. In these matters the operative image of the wasp and the orchid can help identifying

when a digital practice repeat the binary relations I relate to the notion of context. The essential mechanism that the wasp and orchid maintain is that of interface; the cyclic

negotiation by which they perform acts of territorialisation. Similarly I will propose that the medium of the computer must continually be situated and problematized rela-

tive its own externality. We should think of digital practice as always interfaced with a conflicting and ‘dirty’ reality, which does not necessarily comply with the geometric precision and mathematical descriptions of a design space grounded in computation.

As an afterthought this perception would, from a broad perspective, be consistent with

a view expressed by Antoine Picon, where a contemporary technological environment replaces traditional technological objects such as cars and airplanes with more com-

prehensive and abstract entities related to the concepts of networks and fields, which stresses the significance of virtuality (Picon 2010). Accordingly the technological environment has assumed a seamless nature:

“In this environment, components of technological objects are less and less assembled according to schemes based on geometry and mechanics. […] Computers and

more general electronic equipment are no longer designed to these principles. They appear as layered assemblages […] systemic or synergetic organization is replaced

by a different and in some ways looser type of relation based on interfacing. In-

terfacing has more to do with problems of code-writing and translation from one code to another than with traditional structural design.” (Picon 2010, p.120)

228


ENVIRONMENT AND INTERFACE

229


EXPERIMENT 3 | SCATTERINGS


SCATTERINGS

Scatterings developed from a relatively simple interest in the idea that the design of

architectural structures might follow from a condition of provisional relations between

their component parts. In other words, that structural arrangement can emerge independently from a predetermined schema. Though, as the project evolved it also ex-

panded to address questions of decentred control, agency, composite materials and simulation.

In terms of reference I position the work and the final physical demonstrator as a

kind of hybrid practice. This practice could be exemplified by, on the one hand, the controlled, geometric and mathematical hanging chains of Gaudi, and something not

unlike the work of the artist Eva Hesse (1936 -1970) on the other. In contrast to Gaudi the latter deployed her material at the edge of control; productively explor-

ing the instability, unpredictability and unruliness of a material (Fer 2009). I already mentioned Hesse, together with a number of other artists in chapter 6, where Stan Allen describes a noteworthy tendency attached to this group. What is of interest

is that they use material so as to consciously make it impossible to exercise precise control. Instead these artists establish the conditions within which the material will be deployed. They give in to the inherent logic of the material itself and let this shape

and calculate its possible configurations, only loosely orchestrated by the artists themselves (Allen 1997). When I single out Eva Hesse in relation to the present experi-

ment it is because, within her specific practice, her material space of becoming did in numerous projects revolve around the idea of casting and layering in the medium of

rubber, plastic or latex, thus, in a non-industrial way relating to the idea of composite materials. Scatterings explore a similar material context. What I find inspiring in the

work of Hesse is how complexity arises in a material medium apparently outside the possibility of strict mathematical description. Art historian Mark Godfrey describes

Hesse’s work to be more environment than object (Godfrey 2006). Within this en-

vironment Hesse’s methodology is based on construction as a series of linked actions, as opposed to something describable in terms of material treatment. As an example, in the work ‘Right After’ (1969) Hesse combines different kinds of plastic resin and

metal to thematise a dialog between weightlessness and the effect of gravity on the material. Rather than complying with the catenary Hesse counteract it by utilizing the hardening of the resin, to hold irregularities in place, and by the fusioning of parts to

create differentiated material performance. Thus, the memory/time of the hardening process of the resin becomes part of the work.

231


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

[1]

1. ’Right After’ (1969) by Eva Hesse | 2. ‘Untitled (Rope Piece)’ (1970) by Eva Hesse

232

[2]


SCATTERINGS

As a practice - in similarity to the example of Hesse - Scatterings describe an environment in which a material is temporally deployed as a non-representational series of actions at the edge of control, working with and against gravity. Project overview

The project aimed to raise questions concerning the necessity of mapping possible relations between parts of a material structure prior to a process of digital form devel-

opment. Concretely, in the Lamella Flock project I observed a discrepancy between the apparent autonomy of the agent components and the fact that the structure they

sought to form exercised a top-down constrain on their possibilities to connect. In the Porous Ascend project, although generative, an underlying schema directed pos-

sible configurations of components. Could relations be unspecified and provisional

solely forming on the basis of a process of self-organization aimed at a becoming of structure? Through a bespoke digital design system the project explored this possibility

through the idea of assigning a solution-seeking behaviour to representations of the

component parts of a structure being formed. That is, in the Lamella Flock project behaviour was based on knowledge of a possible solution space, whereas Scatterings seeks to define a solution space. Here each component scan and evaluate its surround-

ing in order to determine how to establish connections forming a coherent structure – thus, introducing an investigative agency to the framework of the thesis.

Further into the project this behavioural process would also be constrained by findings

related to the development of physical counterparts to the digital elements. Appearing as clusters of pentagonal transparent patches, these material components were

manufactured as composites consisting of thin acrylic plates and silicone rubber. The combination of these materials followed from a need to achieve a specific material

performance in relation to finalizing the project as a 1:1 spatial demonstrator. This overall performance of the formed structure was explored by developing an additional system for simulating how the structure behaved under gravity.

The project was developed as part of the thesis work but was also used to participate in an exhibition, where it was required to occupy a specific space. This allowed the

work to be directed towards addressing a concrete context and purpose, while retaining enough freedom to stay within the research framework. The title ‘Scatterings’ was

chosen for the slightly ambiguous meaning of this word; on the one hand referring

233


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

to the haphazardly dispersing of physical elements or fragments, while on the other denoting something related to a process of spreading or deflection.

In the following I will describe the two different digital design spaces developed for

the project, the process of creating the physical components and the assemblage of the 1:1 demonstrator.

Digital Design Spaces

The project operated by a division into two different digital design spaces; one for the

interaction of agents forming relations, and one for simulation of the overall structure. Both spaces rely heavily on the computer scientific practice of particle-spring simulation methods.

“Particle-spring systems are based on lumped masses, called particles, which are connected by linear elastic springs. Each spring is assigned a constant axial stiff-

ness, an initial length, and a damping coefficient. Springs generate a force when displaced from their rest length. External forces can be applied to the particles, as in the case of gravitational acceleration. “(Kilian Ochsendorf, 2005)

Each particle element in such a system has a position, a velocity, and a variable mass, as well as a summarized vector unifying the forces acting on it. When not in equilibrium

there will be movement throughout the system as the particles and springs seek their equilibrium positions.

Operationally the particle-spring method was used in three ways. Firstly, in both the

agent-space and the simulation-space the overall shapes of components were defined by particles connected by springs. In the agent part of the modelling this allowed for a

situation where shapes could adapt and transform while remaining able to revert back to the equilibrium initial state of the shape. Secondly, in the agent space the spring

logic could be utilized to form the relations attempting to connect elements; dynami-

cally ‘pulling’ elements together. Thirdly, in the simulation space the description of the structure by means of particle-spring logic allowed simulation of gravitational acceleration and overall tension.

234


SCATTERINGS

Agent Space

The digital design space, in which the agent-based negotiation takes place, consists

of a circular surface on which the component-agents can be dispersed and related. Conceptually this circular surface relates to the idea that the structure evolves as a kind

of unfolded counterpart to its three-dimensional self. The two-dimensionality of the agent-system is consequently observed as a design space abstracted away from a purely

representational outset, towards the idea that it is just as much notational and diagrammatic. Thus, the future process of a transformation towards three-dimensionality

occurs by tensioning the resulting two-dimensional structure of the agents, pulling unconnected elements together, and lifting it off the ground. Therefore, there is no

direct representational similarity between the overall appearance of the structure in its spatial form, and in the agent-space.

In addition to the above, the notational and diagrammatic nature of the surface allowed a mapping of component properties, other than relational, to be assigned to the circular ‘patch’ of the agent-space; namely which sizes and colours components will

assume relative to decisions of the designer based on a reading of the context of the final demonstrator.

Forming Connections

Within the environment of the circular surface, agents commence their process of forming relations by scanning their surroundings for suitable connections to other

components. When they find such a connection the two elements will start to pull towards each other. Since this is a time based process, occurring in a large collec-

tion of elements simultaneously, the different trajectories will start to interfere so that agents shift and loose connections in an on-going choreography until a stabile state is achieved for the system.

The process of establishing connections follows a number of steps for each of the three pentagonal objects forming a component. The first part of this process consists

in separating the candidates for connections from the ones that are unavailable. This unavailability occurs from elements already being occupied, or having a position or orientation, which is not within an elements range of sight.

The next step relates to choosing amongst the candidates which is the most appropriate. Developing this step was noteworthy because it required observation and qualitative

evaluation while writing the algorithm, and could not be acquired by purely analyti-

235


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

[3]

[4]

Distance

Direction

Orientation

3. Examples of agent configurations | 4. Properties evaluated when forming and negotiating connections

236


SCATTERINGS

cal means. That is, to achieve suitable design solutions it was necessary to observe the actual time based behaviour of components in action and reflectively iterate between

these observations and code writing, thus continuing the methodology established

in the Lamella Flock experiment. What was found was that the best behaviour was achieved by weighing different properties of a candidate connection. Obviously the

first of these, which comes to mind, is the distance between components. It is offhand logical to assume that connecting the elements that are closest to each other yields

an immediately equally logical choice. However, it was found that distance alone was

insufficient to produce the most reasonable behaviour. The property of distance needed to be paired with data related to the direction towards the candidate, as well as the

orientation of the edge of the candidate. The reason for this is that, in many situations, the position of two elements might be close to each other but the movement needed to achieve connection requires significant reorientation of the elements. In these cases

following a more direct path turned out to present the best overall solution in terms of stability for the system. In order to refine this relationship between the properties of

distance, direction and orientation the system allows the designer to ‘weigh’ the impor-

tance of each relative to the others. That is, each candidate will be ranged according

to how well they answer to one of the three properties, but their respective positions on these lists would have different weights of importance. This means that the system

might connect two element that are relatively far from each other if the designer has stated that it is directionality or orientation that weighs heavier than distance1. The Makeup of Connections

As already mentioned the technical method for representing relations was a particlespring logic. In this I was interested in developing an alternative approach to that of the Lamella Flock project, in which the algorithmic principles were based on a ‘Boid’-

type logic (Reynolds 1987). The advantage of using particle-spring principles were in this connection that movements as well as individual shapes could be described as a

single system of positions, relations, lengths and tensions. In this way forces and geometry interfaced by means of a shared logic.

When a relation is formed between components the system creates two springs linking the corner particles of the pentagonal shapes to be connected. In less technical 1 In the case of the system for the final demonstrator, from a total of 100% weight, direction was set to 10%, distance 20% and orientation 70%.

237


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

terms, a spring could also be described as a digital elastic band, in this case however with no length. The spring pulls and seeks to achieve its zero length equilibrium. In

this the system produces two kinds of connections, either long ones, when spring equilibrium has not been reached, or tight, when a destination has been arrived at. The

latter are represented simply as two pentagonal elements sharing an edge, while the long connections are formed as a strip flowing from one element to the other. The maximum length of this strip, as well as the threshold distance between elements for forming a connection, was based on fabrication constraints. System Usage

Designing a structure using the agent system is necessarily attached to allowing the autonomous behaviour of the agents play itself out. Nevertheless the designer can take

control over the initial condition of the system, placing agents on the environment

surface, and are allowed the decision to spawn additional agents in the course of run-

ning the system. Additionally it was made possible to place stationary agents within the system – as an example, for the practical purpose of making sure that agents would be present where the overall structure was to be attached to the site.

In addition to placing and initializing agents the user can also place information with-

in the environment affecting the agents. The idea was, similarly to how properties were

orchestrated in Porous Ascend, to project images onto the environment surface and use colour-values to change parameters within the agents, as they travel across this mapping of the environment. This was more concretely used for apportioning the relative sizes of components and their colours.

Simulation Space

Components within the agent system will eventually find an overall steady shape.

This does not necessarily mean that all relations have reached their equilibrium stage, where all connecting edges coincide, but that a balanced tension has been established from which the system does not change. Within the simulation space this stabile

state of the structure is transferred to a three-dimensional space. Here three different

events are simulated. Firstly, how the system reacts under gravity. Secondly, how this plays out in accordance to how the structure is to be physically erected within an ex-

hibition space. Thirdly, how a tensioning of the two-dimensional structure shapes its three-dimensional appearance. Simply put, the system simulates the assemblage and

238


SCATTERINGS

[5]

[6]

5. Screenshots from the ‘agent space’ | 6. Screenshots from the ‘simulation space’

239


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

mounting within the space of the site.

Again this system is based on particle-spring logic, but since it now operates in threedimensions and all elements are mutually interconnected it differs from the agent system in important aspects. The main point of difference is that, in contrast to the agent system, the simulation system will need to address the actual material behaviour of

the physical components. Below I will enter into more details concerning the material nature of the components; however, the main concept behind the material behaviour

is that components and their connections bend. This bending behaviour was imple-

mented within the simulation system by means of additional springs that seeks to pull parts that would otherwise simply hinge towards a straight condition. This allowed for

another kind of tension within the system creating an ability within it to act against the gravitational pull.

Component and Assemblage Appearance

The appearance of the individual components within the structure follows a specific geometric arrangement. Although this arrangement does not change from digital representation to material realization, two different perspectives on the appearance was

used for practical purposes throughout the project; one for the digital processes and one related to fabrication.

As devised within the digital design spaces the individual components appear as a

cluster of three pentagonal figures initially pointing outwards from each other at a 120 degree angle. The three edges facing outwards from the component act as the point of

interface for other components. As described above components can be combined in

two ways; either by means of a tight connection, where edges coincide, or as a loose ‘long’ connection forming an additional connectional strip. This conceptualization was used for both the time based negotiation of the agent system and the simulation.

As the project expanded towards the physical construction of the system the conceptualization of the component required a change of perspective. Consequently, the

appearance of components, as build, became based on how each pentagonal figure had formed a relationship. This meant that a single component in construction consisted

of the two pentagonal figures that formed a connection. This shift was necessitated in order to limit material usage and achieve the strongest joints.

240


SCATTERINGS

[7]

[8]

7 Components in moulds | 8 Axonometric drawing of the component/mould construction

241


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

[9]

9. Final demonstrator as it appeared within the simulation space

242


SCATTERINGS

Materiality

The material development of the components followed from an interest in the idea

of composites. The decision to venture into such a direction was inspired by a reading of the introduction to Gregg Lynn and Mark Foster Gage’s ‘Composites, Surfaces, and Software: High Performance Architecture’ (Lynn Gage, 2011). Here Lynn and Gage propose that new media arises together with changes in concept, construction and

form. They do not lead to, or follow from, the emergence of new formal or constructive paradigms. Lynn and Gage suggest that digital media is in this way inextricably

linked to the idea of composite materials. Thus, according to this view the conceptualizing of composite materials resonates with the utilization of digitally controlled geometries and processes.

In a dialog between the visual appearance of the components and their desired mate-

rial performance the physical components were developed by combining acrylic sheet

material and silicone rubber. Here, individual acrylic pieces were combined by casting them inside a core of silicone rubber. This allowed for a situation where the joining of individual parts was in fact a procedure of merging these parts into a single layered material.

As briefly mentioned in relation to the simulation the use of material was heavily informed by an interest in controlled bending behaviour. In this regard the components

worked with two situations of bending. Firstly, the groups of pentagonal shapes were detailed with a connected overlap of material which allowed the area around their

coinciding edges to open up when a force was applied to them. Secondly, bending was anticipated in the connection between elements. Here it was attempted to limit the

bend in tight connections, by means of overlapping material, and to allow more bending in the longer connections. Fabrication

The moulds used for casting the components were open containers consisting of several different layers tightened together by screws. All parts of the moulds and the

acrylic plates were manufactured by laser cutting. The moulds themselves were made from wood in HDF. Ideally acrylic would have been preferable, due to easier removing

of the finished components, but as a result of economic constraints we did not have this possibility.

243


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

[10]

10. The process of making

244


SCATTERINGS

The overall concept of the moulds was to lock the acrylic plates inside them so as to

allow the liquid silicone to sieve around, under and through them. In this way the hardening process would weave the two plates together. Intensive testing and prototyping was necessary to achieve the desired strength and performance.

The actual process of filling the moulds were done by injecting the silicone by means of syringes through openings in the acrylic plates. This allowed the silicone to fill up

the moulds from bottom to top without significant problems related to the appearance of unwanted air pockets.

When the silicone had dried moulds could be taken apart and the components could be sorted in accordance to an engraved indexing in the acrylic parts. Assemblage

After the casting process had ended the actual assemblage of the final physical demonstrator could proceed. To this end a map of the structure in 2D was used to identify

the position of each element within the structure. Elements where combined in transportable patches and shipped to the exhibition space, where they were assembled in

steps from the perimeter of the structure towards its centre as it was slowly lifted up to its final position.

Findings and Challenges

In the introduction to this text I offered a conceptual framing of the experiment as an environment in which the designer operates at the edge of control and description –

with and in time. Notably, what additionally evolves within this environment does so without an underlying diagram structuring the configuration of elements. This ability lies solely within the agency of the components themselves. Thus, if a diagram does

exist it is represented by the elements themselves. The project stages a situation where mapping does not precede a phase of planning but occurs throughout a process of for-

mation, as performance. Consequently, a specific type of agency develop, which could be denoted as investigative. This concept of an investigative agency, allowing individual

parts of a design to scan their own environment, is one of the key contributions offered by the experiment to the framework of the thesis.

As within the previously described experiments Scatterings also direct attention to-

wards new problems, unveiled by how the experiment unfolded and was concluded. The main difficulty suggested by the experiment relates to the division into two sepa-

245


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

[11]

11 Close-up of the demonstrator

246


SCATTERINGS

rate digital design spaces, and a question of the limit to what can be automated and related.

The method of negotiating the configuration of the components depends on being

played out within an abstract two-dimensional space, which does not communicate reflexively with the three-dimensional space of the simulation. Since this space depends

on a coherent configuration of elements, the process of mapping the components is terminated when it is transferred to the simulation space. In this way a bottleneck occurs by which observations made within the simulation space – requiring alterations

to the initial configuration – necessitate that the negotiation of the agents will have to be started anew. On the one hand this suggests that it would be preferable to increase

feedback between simulation and mapping, however, on the other hand the question is how to utilize such a feedback. What I observe is that this would be limited to

what can be described in terms of quantifiable parameters alone, while the qualitative evaluation of a design will remain within a domain of description that is necessarily

vague, open to interpretation and intuitive. Thus, increased feedback could offer an optimisation of workflow, which would be meaningful in the context of general purpose design systems. However, when we deal with project-specific systems, as in the case of all experiments attached to the thesis, the effort required in order to implement a

higher degree of fluidity and automation should be weighed against the benefit and

cost to the project at hand. The framework for design suggested by the thesis does not

attempt to insert systems that are general, but ones that are specific to the context of a project. Here individual algorithmic techniques, defining generative processes, are

tools yielded, altered and chosen by the designer. By being part of a process of design

they refer to a qualitative complexity identified with the agency of the designer. Thus, the experiment exemplifies a need to consider, as part of a digital practice, how far one

need to go in an effort to generate consistency across the elements of a design space. Certainly, more consistency would have been preferable within Scatterings, but, what

is more important to note is that within a digital design practice priorities are required. My suggestion in this regard is to emphasise the distinct and unique complexities of

the digital and the cognitive capabilities of the designer, and to deprioritise instrumental efficiency, since we are dealing with unique design processes and not software development.

247


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

[12]

12. The final demonstrator | 13. Close-up of the demonstrator

248

[13]


SCATTERINGS

[14]

[15]

14.-15. Experimental probe. System created for the studying of agent-based principles for organising the assemblage of structures in three dimensions. The box-like objects follow an investigative agency through which they negotiate and compete in order to form a stable configuration.

249


ENDING



8 | FRAMEWORK


FRAMEWORK

I perceive the overall outcome of the thesis to assume two forms relative to how it ap-

pears within the terrain established by the present work – a form of solidification and a form of continuing movements within this terrain. The first type of outcome relates

to momentary generalisations to be drawn from the work, the other to the manifesta-

tion of a productive surface defining a territory of further investigation. This surface is ultimately a product of dynamics and interrelationships across the generalisations, and as such the two forms of outcome are necessarily interrelated.

This concluding chapter describes two corresponding outlines originating within

these observations. Firstly, I will discuss a set of practice related proposals discovered

throughout the thesis, which are directed towards a more general context of experimentation by means of generative processes. Secondly, I will turn towards an internal reflection across the productive zones and diagonals within the present material. This

latter ultimately allows for an understanding of how a design process might unfold across the resulting complete territory of the thesis – forming a position from which new investigations might commence. Summarising the Problem Domain

The thesis constructs a series of elements, which in concert are understood to define a

map constituting a framework for a specific type of design practice. This practice departs from a critical exploration of potentials associated with the use of computational

technology, with the specific focus of understanding generative processes, as meaningful constituents of a design process. Relative to this point of departure the thesis has

developed from a particular technological problem domain observant of a core issue related to the mechanism of reflexivity.

In the introductory chapter I identified this problem domain to proceed from an ambition to understand a condition of a difference-making conflict – rather than opposition – between, on the one hand, the planned, controlled and hierarchical, and on the other, the self-organising, non-linear and autonomous. Here the mechanism

of reflexivity has been developed as a potential technological hinge between these ap-

parent opposites. The proposed significance of this concept is that it can be used in an effort to create an operational space, which recognises the autonomous potential of the generative, while acknowledging that for its corresponding processes to be meaning-

ful in design, unitary events – imposing their authority on the system – are required. That is to say: a generative system is initially set in motion by a designer, placed in

253


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

an appropriate environment; its goals are not specified by itself; and its end-result must necessarily constitute a degree of unity. Simultaneously, in the intermediate space between these events the system does unfold autonomously and thus makes these events

part of an iterative process. Herby, the elements which generate the system are made to

become part of it. This creates a condition within the design space, which I described in the introduction as a loss of hierarchy; a concept that should be understood in the sense

that authorship and authority, within the system, is fleeting, changing and unstable – however still existing – creating a dynamics of difference.

Solidifications: Propositions and Generalisations

I will in the following section outline the overall propositions that the thesis makes

for an experimental architectural practice engaged in the use of generative processes, observant of how these proposals address the problem domain defined by the core issue

of reflexivity. These proposals embody the broadest generalisations to be drawn from the

thesis; they are made as a suggestion, offered more as a supplement than as a replacement of current design strategies. My intention is not prescriptive. 1st Proposition: An approach to Technology

The application of a generatively situated technology extends beyond the scope of an instrumental reason and an applied science

The model of the framework designates intelligence to be distributed and it questions a

clearly defined position of authorship. It subscribes to a view of nature as technologized,

where reason and agency exist within the object and is not solely imposed upon it. Thus, practice never occurs from an outside, but is conducted as a locally determined agency amongst a multiplicity of other agencies. Representation is replaced by reflexivity and

planning becomes generative and algorithmic. The design space assumes the form of an evolving system without exteriority in which observation always includes participation.

By operating relative to the mechanism of reflexivity technology cannot comply with the representational knowledge ascribed to an institutionalised science, since it is not applied

as a means to predict or plan a process of design. The concept of technology, within the framework, is defined as different from science, and is thus not limited to operate by its

mechanisms alone. This allows technology to also be directed by the dynamics of qualitative differences.

254


FRAMEWORK

2nd Propositon: Digital Practice

By conceptualising a larger field of digital practice as approachable by means of generative

techniques and the mechanism of reflexivity, the proposed framework for design constitutes a

surface for experimentation across varying tendencies characterising a contemporary digital architecture

This proposal is not to be understood as an attempt to unify the field of a digital archi-

tecture, but an indication of the framework as a possible experimental and exploratory

practice within, and contributing to, a larger field of practices. It is the result of presuming that contemporary digital practices relate to a turbulent condition in which different

emphases of a digitally produced architecture mix, conflict and evolve together. In this process I observe generative aspects to flow across other current focuses. Here, explora-

tions of material concerns are seen to perceive material behaviour and characteristics as an agency assisting in coupling architecture to its environment; complex informational

modelling result in design spaces characterised by non-linear interdisciplinary systems of internal exchange; and cultural intensities are interfaced reflexively with the application of novel technologies.

From here, the practice of the framework is proposed to be applicable as an operational tool, which allows us to interface with a wider range of contemporary concerns, making

these internal to it. Thus, relative to the framework, generative practice combines with the material, informational and cultural, making the concept of reflexivity the overall operational mechanism by which to observe their corresponding technologies. 3rd Proposition: Design Approach

In order to produce compatibility between the writing of generative computational processes

and design practice, the digital design space should be understood as a complex system of learn-

ing and observing, where notions of planning and representation are substituted by reflexivity and performance

The design approach suggested by the framework designates a notion of the design

space as a non-linear system of learning and observing, where design solutions are arrived at by way of a decentred agency existing between its constituents – including the designer. With this approach to design, a process of planning – and the ideals of an

absolute representational knowledge and complete control – gives way to a reflexive

255


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

process of indeterminacy and performance. This is seen to apply to design in general but specifically observed in relation to generative design processes in which coding/ programming is applied.

With this approach design solutions are arrived at, neither through the agency of the model, nor that of the designer, but through a decentred reflexive agency. Here the

role of the designer turns from being that of a predicting and imposing agent, to an observing and learning mediator. In this process the designer seeks a mode of enabling control, rather than restrictive; something which allows us to steer relative to complexity without reductionism. This also implies that the designer becomes positioned

within the system of design, as opposed to being the observer of the system. Thus, the controller and what is controlled are related by circular causality. The designer steers

within the system; all the while the system itself assumes an agency by which it exercises a degree of self-control.

Through this proposal the design space assumes the form of a system that the designer

– seen as an adaptive agent – performatively interacts with in an open-ended exchange to produce difference. Here the concept of reflexivity can be understood as the necessary operational mechanism by which change occurs. 4rd Proposition: Representation

The notion of agency can be conceptualizes as a mechanism introduced into representation The specific approach to design suggested by the framework problematizes the con-

cept of an eidetic and absolute form of representation; this, because it subscribes to a

perception of reality founded on the performativity ascribed to the concept of agency, which opposes the notion of prediction implied by resemblance. Nevertheless, architectural practice is inherently tied to representation, wherefore I suggest that repre-

sentation can be conceptualised as a vehicle for the agency implied by a generative perspective on design. Consequently, within the framework for design, I propose to

envision what constitutes representation as a device for steering within the space between immaterial design development and physical making, in the form of a map comprised of concrete form and abstract notation. The notational aspect allows the

map to project agency by way of an intensive operationality from which a reflexive relationship to the build might be established.

What constitutes representation within the framework consist of two components.

256


FRAMEWORK

A cartography related to a gathering of both agency and the forms to which it is at-

tached, and a field condition which supplies a mode by which to think the exercise of

agency; i.e. the agency attached to a design. The cartographic approach to representation proposes the concept of the map as a way to think the medium, and it establishes the design space as a landscape in which something can be discovered – and eventu-

ally created. The mapping allows us to define and describe local conditions relative to which we may deploy a process of design. The map represents a field condition. By

making this field digital its agencies may be simulated. Architectural representation becomes a record of the processes made possible by the interacting agencies within

their environment. We achieve a non-hierarchical horizontality, which by addressing and staging agency allows it to become a tool by which to perform through, as opposed to one of planning and prediction. 5th Proposition: Design Space

The model of a design space should be conceptualized relative to a notion of environment The concept of environment represents a diagrammatic and operational comprehension of the design space, which designates a responsive, conditional and complex

entity. The concept of environment does not reference an exteriority, but a dynamic space both formed by and forming a process of design, thus resisting a restrictive

binary specification of process and absolute representation. The appearance of my

model in this regard is that of a double articulation, its structure is algorithmic and its overall being relates to a condition of virtuality.

In my understanding of the concept of environment this entity constitutes a situa-

tion where a duality between object and exterior dissolves in a doubly projective and reflexive movement. Here the boundaries between forms and forms, and surround-

ings and form are not clearly distinguishable. A folding occurs that makes surroundings (including tools and technologies) immanent to the design space and what it produces. The concept of environment substitutes that of the object, in the sense

that a form becomes a feature within the environment. Through this characteristic it becomes necessary to always think the medium of the computer as situated and

problematized relative to the larger space it inhabits. That is, the medium is itself a

feature of a responsive, conditional and complex environment, which replaces binary organisation with relations based on interfacing.

257


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

To summarise: The thesis proposes a framework in which a generative practice

navigates relative to five stabile points of orientation. 1. A technology maintaining the constant of change, operating beyond an instrumental reason. 2. A wider con-

temporary field of practices reflexively interfacing a generative approach to design

with material, informational and cultural dynamics. 3. A view of the designer as an

embedded observer reflexively seeking an enabling form of control within the design environment. 4. A form of representation envisioned as a cartographic device for

steering relative to the agency of the generative design space. 5. A conceptualisation

of the design space as a doubly-projective environment dissolving the divide between surroundings and object.

Currents: A Territory of Further Investigation

The above proposals for a generative practice denote a degree of solidification within the problem domain observed by the thesis. They are momentary points of orientation, which can serve to relate and contrast the work to a larger field of current

practices, and a measure of the present status of the research. In this regard I perceive the outcome of the thesis to indicate a potential for further investigations. From a

perspective of continuity the significant element is therefore not the set of unitary

components as such, but that of the territory forming between and relative to these solidifications, since this space necessarily defines the continuous movements suggested by the thesis. This is where creative disturbances may be found, where new

solidifications might appear and current ones dissolve. Consequently, I perceive the

result of the thesis as a relational map, designating both productive zones and unproductive channels, meetings and fractures, which does not prescribe a next project as such, but represents something strategically applicable in such an endeavour.

A Practice Aligned to the Framework

Central to this thesis has been a desire to develop an understanding of the processes

by which digital tools and media operate relative to a generative approach to design.

Specifically I aimed towards addressing this topic in a movement that simultaneously acknowledges and ventures beyond the technical execution, in order to explore how the processes in question might encourage new ways of thinking a design process.

With this outset I sought to form a conception of how the formulation of algorithms and computational processes might serve a reflexive coupling to already existing sys-

258


FRAMEWORK

tems of a complex surrounding reality, resulting in a fabrication of difference within a productive conflict between top-down and bottom-up technological processes. From the perspective of my own practice – as an architectural researcher writing

computational processes – a surveying across the territory of the thesis now allows me to make suggestions as to the nature of such an understanding. Here I depart

from a notion of design practice as something which must unfold within a produc-

tively underspecified space, where planning gives way to processes of indeterminacy

and reflexive performance – that is, the space in which the solidifications of the pre-

vious section allow me to navigate. I will in the following venture across the territory forming in-between these unitary elements, in an effort to contemplate how digital

design practice might unfold in this setting. Not as an attempt to make a generalis-

ing statement as to how such a design practice should unfold, but to establish a space which can act as a starting point for new explorations. In this way the following

should not be read as the account of a methodology, but as principles for experimentation through a generative design practice operating by means of computationally

articulated processes. Also, the account represents a synthesis between findings within the modes of discourse and making by which the thesis has been developed. Here the conceptual and speculative encounters the procedures and processes explored through the technical. Field

The framework resulting from the thesis proposes a design space, which displays a

perception of a surrounding reality that is not susceptible to an absolute representation and cannot be a target of an instrumental reason. The world is here observed

as a system filled with agency, relative to which we act from within, by discovering

– rather than following – rules. In this context there is no exteriority from which a

design problem can be stated and exhaustively specified, because such a world cannot be enframed. With this perception of reality, design is suggested to commence

by means of a cartographic approach, where agencies and forms – to which we seek to relate a design process – are tactically extracted forming an associated environ-

ment. This designates a first requirement to be followed by an algorithmically driven

practice of design, since the processes this type of practice specifies should be able to

mediate the notational and diagrammatic aspects of the cartography. That is, not only proceeding relative to spatial forms, but relative to simulations of the agencies gath-

259


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

ered. Here the object can only be fully approached from the depths of algorithmic

generation and the design space is never a target or tool for rationalisation. Technology is applied as a means to maintain the constant of change in varying and evolving environments. In this way a generative understanding of design becomes more

than an instrumental approach to the making of an architectural artefact, in that it additionally forms the basis of a mediation of reality, which itself is perceived as a

generative system. The map becomes a way of thinking the computational medium – incorporating the design space as a multi-scalar and evolving landscape in which something can be discovered, and eventually created by means of a reflexive move-

ment. Here the complexity of the object is replaced by the complexity of the space in which the object evolves. This observation can be of use when contemplating the use and nature of algorithms in the present practice. These are to be understood as context-dependant mechanisms, conditioned by a dependency on the environment in

which they are articulated; something which has an effect on a design process, while not resembling the result of this process. By way of these entities the expression of form is always a feature of, and within, an environment.

Through a cartographic approach to the establishment of an appropriate design

environment, a situation is arrived at where the success and productivity of such an environment depends on its ability to establish a point of departure characterised by a productive indeterminacy – a condition of difference capable of diversifying

the range of possibilities a process of design might pursue as it propagates from the

surface of mapping to a construction of an architectural artefact. To commence this

movement a design process departing from computational code-writing enters into a second stage. While the first stage (that of the cartography) establishes an algorithmic interface to the given, this second utilises the cartographic surface as a means to deploy an agency attached to the new. This is the point where the design space

projects back into a surrounding reality, as mediated by the computer. The second stage extracts, and is triggered by, data from the cartography, and as such exhibit

similarities to the process of mapping, however what is specifically characteristic of its procedures is that they involve the invention, introduction and discovery of new components and forms in relation to an associated design environment.

By projecting the agency-of-the-new onto the agency-of-the-given an iterative pro-

cess commences, which, from the perspective of the design space, in a sense results in a cancelling out of this binary schema. A blurring of boundaries between a real

260


FRAMEWORK

and a constructed world takes place, offering primacy to neither. Here the architectural model exhibit a plasticity relative to both its immediate inherent organisation (parameters and constraints) and an externally specified temporality, which allows

for creative instabilities and discontinuities to sieve in from its evolving associated

environment. In this process the designer assumes the role of an active and involved observer, both in the process of designing the system, and designing with it. In this way a methodological model for design is assumed, which has its procedural basis

within the locally determined, autonomous and individual, characterised by a process of iteratively coding, manipulating, adding information and observing what occurs

within the simulated dynamics of the differences existing between the orchestrated agencies. The activity of computational coding is therefore only part of the process, which extends beyond this practice; into heuristics, material testing, information

gathering, aesthetics and prototyping. The facilitating mechanisms driving a system for design must lie outside the technological. But, since this outside is beyond an absolute representation and uniquely associated to individual design problems it

becomes overall futile for an architectural designer to proceed by means of developing generalizable computational tools and techniques. These are necessarily always

bespoke and a product of the cartography. This implies a continuous iterative pro-

cess of developing the cartography and altering the components/entities, algorithms and processes deployed within the field it establishes. The aim is to define a solution space from within, through the agency of its components and the positionality of

the designer. From here design operates without an underlying diagram – instead it

forms it. In accordance to the non-representational perception of reality, this diagram references no external coordinization but develops by means of both temporal and

spatial parameters within the system, pointing towards a reactive and adaptive organization of architectural elements within the field described by the cartography.

In this iterative process the aim is to establish compatibilities across territories, e.g. between technology and material. Since the design space is not perceived as a target of rationalisation the designer proceed in these matters by striving for a mode of enabling

control, as opposed to a restrictive. The ideal is to establish a system where informa-

tion and control-actions allow a practice of design to steer towards and maintain goals, while adapting to the agency of an associated, dynamic and layered environment. Here

the entities of the design process become adaptive devices within the system of the computational medium. The challenge in these matters is to create the necessary abil-

261


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

ity of interfacing between the constituents of a total process of design, while balancing between the homogenisation that this requires, and the variety and heterogeneity essential to the success of a process of design.

With this process the aim is to think beyond the hierarchy of a goal. Instead the

ideal is to establish the architectural artefact as a function that stabilises on particular values derived from the heterogeneous agencies comprising the design system. This event occurs when the system is not exposed to additional disturbances from

the projection – by the cartography – of additional external relations. Consequently, the finished design appears as something that comes into being within the design space and at a certain point persist within a particular state. Thus, the purpose of

developing a computational system for a design project, as envisioned here, is to create the conditions in which the design outcome can come into being and continue

to generate itself. The design space is perceived as a flow which solidifies into unities but always arrives from a structure of multiplicity. In these matters we can return to

the significance of the underspecified character by which I earlier denoted the design space. Within these systems the ideal should be to operate outside prediction at a

complexity where variety is beyond absolute representation. Here a necessary loss of top-down control allows for a staging of the design problem as unmanageable, but within this unmanageability we encounter events of discovery. It is through these

events that design as planning gives way to design as a reflexive activity. Since reflexivity here is operational through technology, I will claim that the proposed design

process significantly define difference as something that can never be produced by

technology, but only in an external relation to it – within an environment. Technology does not embody difference but is a tool by which to reveal it.

Closing Remarks

I believe that the terrain of the proposed framework unavoidably points towards its own future development. This appears through an inherent potential and necessity

within the above conception of a generative design process to always proceeding by applying it to new research questions and design problems. It is only truly operational

in such a relation. I believe that the contribution made here, precisely because of its form as a framework, is something which must be projected onto new territories and situations. In this way the thesis is seen to represent an outcome designating a posi-

tion from which new departures for a digitally produced architecture can be made.

262


FRAMEWORK

As stated in the very beginning of the thesis, my interest in this type of architectural

practice follows from a fascination of its continuing process of evolving into yet more mature and complex forms. Seen as part of such a trajectory, future iterations of the

present body of work might proceed by a movement away from the context of the laboratory situation towards concerns related to the scale and complexity of genuine

architectural composition. This would inevitably result in productive encounters and

conflicts, to which the thesis’ key concepts of the generative, reflexive and technological could be challenged anew. It is because of the possibility of such a situation that

these concepts, from my perspective, appear so appealing – as productive influences within any environment to which they might be attached.

263


BIBLIOGRAPHY


BIBLIOGRAPHY

Adorno, T.W., 1958. The Essay as Form. Translated from German by Hullot-Kentor, B. & Willin, F., 1984., in New German Critique, No. 32, pp.151-171.

Alexander, C., 1964. Notes on the Synthesis of Form. Cambridge: Harvard University Press. Alexander, C., 1971. The state of the art in design methods. DMG Newsletter, 5 (3), pp.3-7.

Allen, J.S., 1999. A Short History of Lamella Roof Construction. Transactions of the Newcomen Society, Vol 71, No 1.

Allen, S., 1997. From Object to Field. In: Lynn, G. ed., AD Architecture after Geometry, Profile No. 127, London: John Wiley & Sons Limited.

Allen, S., 2000. Introduction: Practice vs. Project. In: Allen, S. ed., Practice: Architecture, Technique and Representation. Amsterdam: OPA.

Allen, S., 2009. Notations + Diagrams: Mapping the Intangible. In: Allen, S. ed., Practice: Architecture, Technique + Representation. New York: Routledge.

Aranda, B. & Lasch, C., 2006. Pamphlet Architecture 27: Tooling. New York: Princeton Architectural Press. Aranda, B. & Lasch, C., 2008. PIN-UP magazine, Issue 4.

Archer, B., 1995. The Nature of Research. Co-design, interdisciplinary journal of design, January 1995, pp.6-13.

Ashby, W.R., 1956. Introduction to Cybernetics. London: Chapman and Hall.

Ashby, W.R., 1960. Design for a Brain: The Origin of Adaptive Behaviour. 2nd ed. London: Chapman & Hall. Bachelard, G., 1969. The Poetics of Space. Boston: Beacon Press.

Beim, A. & Thomsen, M.R., 2011. The Role of Material Evidence in Architectural Research: Drawings, Mod-

els, Experiments. Copenhagen: The Royal Danish Academy of Fine Arts, Schools of Architecture, Design and Conservation.

Beer, S., 1959. Cybernetics and Management. London: The English Universities Press. Bergson, H., 1998. Creative Evolution, Dover Publications.

Bertram, P., 2011. Evidence of Material Difference. In: Beim, A. & Thomsen, M.R., 2011. The Role of

Material Evidence in Architectural Research: Drawings, Models, Experiments. Copenhagen: The Royal Danish Academy of Fine Arts, Schools of Architecture, Design and Conservation, pp.38-49.

Bonabeau, E., Dorigo, M. & Theraulaz, G., 1999. Swarm Intelligence: From Natural to Artificial Systems. New York: Oxford University Press.

Bonta, M., Protevi, J., 2004. Deleuze And Geophilosophy: A Guide And Glossary. Edinburgh: Edinburgh University Press.

Carranza, P.M. & Coates, P., 2000. Swarm modeling: The use of Swarm Intelligence to generate architectural form. Proceedings of the International Conference on Generative Art.

Cartwright, N., 1999. Models and the limits of theory: quantum Hamiltonians and the BCS model of superconductivity. In: Morgan, M.S. & Morrison, M., eds. Models as Mediators – Perspectives on Natural and Social Science. Cambridge: Cambridge University Press.

Cilliers, P., 1998. Complexity and Postmodernism: Understanding complex systems. New York: Routledge.

Corbellini, G., 2009. Bioreboot: The Architecture of R&Sie[n]. New York: Princeton Architectural Press.

Corner, J., 1999. The Agency of Mapping: Speculation, Critique and Invention. In: Cosgrove, D., ed. Mappings. London: Reaktion Books.

Cross, N., 2007. Editorial: forty years of design research. Design Studies Vol 26 No 1, pp. 1 – 4.

265


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

Delanda, M., 2005. Intensive Science and Virtual Philosophy. London: Continuum.

Delanda, M., 2007. Deleuze: The Use of the Genetic Algorithm. In: In: Braham, W.W. & Hale, J.A. eds. Rethinking technology, A reader in architectural theory. London/New York: Routledge.

Deleuze, G. & Guattari, F., 1987. A Thousand Plateaus: Capitalism and Schizophrenia. Minneapolis: University of Minnesota Press.

Deleuze, G., 2004. Focault. Copenhagen: Det lille forlag.

Dawkins, R., 1976. Hierarchical organization: a candidate principle for ethology. In: Bateson, P.P.G. &

Hinde, R.A., eds. Growing points in ethology : based on a conference sponsored by St. John’s College and King’s College. Cambridge: Cambridge University Press.

Dorin A. et al., 2012. A Framework for Understanding Generative Art. Digital Creativity, Vol.23(3-4), London: Routledge, pp.239-259.

Elkins, J., 2009. On beyond Research and New Knowledge. In Elkins, J. ed. Artists with PhDs: On the New Doctoral Degree in Studio Art. Washington: New Academia Publishing, pp.111-133.

Evans, R., 1997a. Translations from Drawing to Building. In: Johnston, P. ed. Translations from Drawing to Building and Other Essays. Cambridge: MIT Press

Evans, R., 1997b. The Developed Surface. In: Johnston, P. ed. Translations from Drawing to Building and Other Essays. Cambridge: MIT Press.

Fer, B., 2009. Eva Hesse: Studiowork. New Haven: Yale University Press. Feyerabend, P., 2010. Against Method. Brooklyn: Verso.

Flake, G.W., 1998. The Computational Beauty of Nature: Computer Explorations of fractals, chaos, complex systems, and adaption. Cambridge: MIT Press.

Forty, A., 2000. Words and Buildings: A Vocabulary of Modern Architecture. London: Thames and Hudson.

Galanter, P., 2003. What is generative art? Complexity theory as a context for art theory. Proceedings of the 6th international conference: Generative Art 2003, Milan.

Gedenryd, H., 1998. How Designers Work. Making Sense of Authentic Cognitive Activities. Lund: Lund University.

Gibson, W.J., 1986. The Ecological Approach to Visual Perception. Psychology Press.

Giere, R.N., 1997. Understanding Scientific Reasoning. Fort Worth: Harcourt Brace College Publishers

Glanville, R., 2007. Try again. Fail Again. Fail better: the cybernetics in design and the design in cybernetics. In Glanville, R. ed. Kybernetes – The international journal of cybernetics, systems, and management sciences, Vol.36, Number 9/10, Emerald Group Publishing, pp.1173-1206.

Godfrey, M., 2006. A String of Nots: Eva Hesse’s Hanging Sculpture. In: Sussman, E. & Wasserman, F. eds. Eva Hesse: Sculpture. New Haven: Yale University Press.

Goodman, N., 1976. Languages of Art: An Approach to a Theory of Symbols. 2nd ed. Indianapolis: Hackett Publishing Company.

Goulthorpe, M., 2008. Decoi Architects: From Autoplastic to Alloplastic. Editions HYX

Grobman, Y.J. & Neuman, E., 2012. Performalism – Form and Performance in Digital Architecture. New York: Routledge.

Gros, C., 2008. Complex and Adaptive Dynamical Systems. Heidelberg : Springer.

Hardt, M. & Negri, A., 2009. Commonwealth. Cambridge: Harvard University Press.

Hayles, N.K., 1999. How We Became Posthuman. Chicago: The University of Chicago Press.

Hensel, M., Menges, A. & Weinstock, M., 2004. Frei Otto in Conversation with the Emergence Design

266


BIBLIOGRAPHY

Group. In: Hensel, M., Menges, A. & Weinstock, M. eds. AD Emergence: Morphogenetic Design Strategies. London: John Wiley & Sons.

Hensel, M. & Menges, A., 2007. Morph-Ecologies: Towards Heterogeneous Space In Architecture. London: AA Publications.

Hensel M. & Sunguroglu, D., 2008. Material Performance. AD Versatility and Vicissitude: Performance in Morpho-Ecological Design, Profil No. 192. London: John Wiley & Sons.

Hensel, M., Hight, C. & Menges, A., 2009. En route: Towards a Discourse in Heterogeneous Space

beyond Modernist Space – Time and Post-Modernist Social Geography. In: Hensel, M., Hight, C. & Menges, A. eds. Space Reader – Heterogeneous Space in Architecture (AD Reader). London: John Wiley & Sons

Hensel, M. & Menges, A., 2009. The Heterogeneous Space of Morpho-Ecologies. In: Hensel, M.,

Hight, C. & Menges, A. eds. Space Reader – Heterogeneous Space in Architecture (AD Reader). Lon-

don: John Wiley & Sons

Heylighen, F. & Joslyn, C., 2001. Cybernetics and Second-Order Cybernetics. In: Meyers, R.A. ed. Encyclopaedia of Physical Science & Technology. 3rd ed. Academic Press.

Holzner, H., 1999. Entwicklung eines Nachweisverfahrens zur Bemessung von speziellen (maschinell gefertigten) Zapfenverbindungen. Diploma thesis at ‘Institut für Tragwerksbau - Fachgebiet Holzbau’, TU Munich.

Johnson, P.A., 1994. The Theory of Architecture: Concepts, Themes & Practices. New York: Van Norstrand Reinhold.

Jones, J.C., 1970. Design methods. New York: Van Nostrand Reinhold.

Kiesler, F.J., 2007. On Correalism and Biotechnique. In: Braham, W.W. & Hale, J.A. eds. Rethinking technology, A reader in architectural theory. London/New York: Routledge .

Kilian, A. & Ochsendorf J., 2005. Particle-Spring Systems for Structural Form Finding. Journal of the International Association for Shell and Spatial Structures: IASS, vol.46, pp.77-84.

Klinger, K., 2008. Relations: Information Exchange in Designing and Making Architecture. In Kolarevic B. & Klinger, K. eds. Manufactoring Material Effects: Rethinking Design and Making in Archtiecture. New York: Routledge.

Kolarevic, B., 2003. Architecture in the Digital Age – Design and Manufacturing. New York: Spon Press.

Kotnik, T. & Weinstock, M., 2012. Material, Form and Force. In: Menges, A. ed. Architectural Design 02 2012 – Material Computation. London: John Wiley & Sons.

Kwinter, S., 2001. Architectures of Time. Cambridge: MIT Press.

Lash, S., 2007. Intensive Media: Modernity and Algorithm. In: Shamiyeh, M. ed. Organizing for Change: Intergrating Architectural Thinking in Other Fields. Basel: Birkhäuser.

Lash, S., 2010. Intensive Culture: Social Theory, Religion and Contemporary Capitalism. London: Sage Publications

Leach, N., 2004. Swarm Tectonics. In: Leach, N., Turnbull, D. & Williams, C. eds. Digital Techtonics. Chichester: Wiley Academy.

Lynn, G. ed., 1993. Architectural Design: Folding in Architecture, London: John Wiley & Sons

Lynn, G., 1998. The Folded, the Pliant and the Supple. In: Lynn, G. ed. Folds, Bodies & Blobs: Collected Essays. Brussels: La Lettre Volée.

Lynn, G., 1999. Animate Form. New York: Princeton Architectural Press.

267


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

Lynn, G. & Gage, M.F., 2011. Composites, Surfaces, and Software: High Performance Architecture. New Haven: Yale School of Architecture.

Lyotard, J.F., 1984. The Postmodern Condition – A Report on Knowledge. Translated from French by Bennington, G. & Massumi, B. Manchester: Manchester University Press.

Mathews, S., 2007. From Agit-Prop to Free Space: The Architecture of Cedric Price. London: Black Dog Publishing.

Maturana, H. & Varela, F., 1980. Autopoiesis and Cognition: The Ralization of the Living. London: D.Reidel Publishing Company.

McCullough, M., 1998. Abstracting Craft. Cambridge: MIT Press.

Menges, A., 2004. Morpho-Ecologies: Approaching Complex Environments. AD Emergence: Morphogenetic Design Strategies. London: Wiley-Academy.

Menges, A., 2012. Material Computation: Higher Integration in Morphogenetic Design. Architectural Design, 82, Issue 2. London: Wiley-Academy. pp.14-21.

Mattelmäki, T., 2006. Design Probes. Helsinki: The University of Art and Design Helsinki.

Pask, G., 1969. The Architectural Relevance of Cybernetics. Architectural Design, 39, pp.494 – 496.

Pask, G., 1971. A Comment, a Case History and a Plan. In: Reichardt, ed. Cybernetics, Art, and Ideas. New York: New York Graphic Society, pp.76-99.

Pask, G., 1975. Conversation Theory. London: Hutchington.

Penrose, R., 1974. The Role of Aesthetics in Pure and Applied Mathematical Research. Bulletin of the Institute of Mathematics and Its Applications, Vol. 10, pp.266-71.

Pickering, A., 1995. The Mangle of Practice: Time, Agency and Science. Chicago: The University of Chicago Press.

Pickering, A., 2008. New Ontologies. In Pickering, A. & Guzik, K. eds. The Mangle in Practice. Science, Society and Becoming. Durham and London: Duke University Press, pp.1-14.

Pickering, A., 2010. The Cybernetic Brain. Chicago/London: The University of Chicago Press. Picon, A., 2010. Digital Culture in Architecture. Basel: Birkhäuser.

Popovich, O., 2008. Reciprocal Frame Architecture. London: Architectural Press.

Prigogine, I., 1980. From Being to Becoming: Time and Complexity in the Physical Sciences. New York: W.H.Freeman & Company.

Rahim, A., 2006. Catalytic Formations: Architecture and Digital Design. New York: Taylor and Francis.

Ralph, P. & Wand, Y., 2009. A Proposal for a Formal Definition of the Design Concept. In: Lyytinen, K., Loucopoulos, P. & Mylopoulos, J., Robinson, B. eds. Design Requirements Engineering: A Ten-year Perspective. Berlin: Springer-Verlag, pp.103-136.

Reas, C. & McWilliams, C., 2010. Form + Code: in design, art, and architecture. New York: Princeton Architectural Press.

Reiser, J. & Umemoto, N., 2006. Atlas of Novel Tectonics. New York: Princeton Architectural Press.

Reynolds, C.W., 1987. Flocks, Herds and Schools: a distributed behavioral model. Computer Graphics, 21(4), pp.25-34.

Roche, F., 2012. Reclaim Resi[lience]stance//….R2. Log 25. New York: Anyone Corporation. Schindler, C., 2009. Genagelt und geschraubt. ARCH+ 193, Aachen: ARCH+ Verlag.

Schumacher, P., 2011. The Autopoiesis of Architecture: Vol.1, A New Framework for Architecture. Chichester: John Wiley & Sons.

268


BIBLIOGRAPHY

Schumacher, P., 2012. The Autopoiesis of Architecture: Vol.2, A New Agenda for Architecture. Chichester: John Wiley & Sons.

Serres, M., 1995. Genesis. Ann Arbor: The University of Michigan Press.

Shelden, D.R. & Kashyap, S., 2012. Gehry Partners, LLP/Gehry Technologies. In: Grobman, Y. &

Neuman, E. eds. Performalism – Form and Performance in Digital Architecture. New York: Routledge.

Simon, H., 1996. Science of the Artificial. Cambridge: MIT Press.

Spuybroek, L., 2004. NOX: machining architecture. London: Thames & Hudson.

Stam, J., 2003. Real-Time Fluid Dynamics for Games. Proceedings of the Game Developer Conference, San Jose.

Tamke, M., Thomsen, M.T. & Riiber, J., 2008. Complex Geometries in Wood. In: Advances in Architectural Geometry 2008. Wien: Springer.

Tamke, M. & Thomsen, M.R., 2009. Implementing digital crafting: developing: it’s a SMALL world. Proceedings of the Conference: Design Modeling Symposium Berlin 2009, pp. 321-329.

Terzidis, K., 2006. Algorithmic Architecture. Oxford: Elsevier.

Thomsen, M.R. & Tamke, M., 2009. Narratives of Making: thinking practice led research in architecture.

Proceedings of Communicating by Design, International Conference on Research and Practice in Architecture and Design, Bruxelles.

Thomsen, M.R., 2011. Computing the Real: Time, Scale and Material. Copenhagen: The Royal Danish Academy of Fine Arts, School of Architecture.

Thomsen, M.R. & Tamke, M., 2012. The Active Model: A calibration of material intent. In Ayres, P. ed. Persistent Modelling: Extending the role of architectural representation. New York: Routledge.

Thompson, D.W., 2010. On Growth and Form. Cambridge: Cambridge University Press.

Tristan, S., Martin, S. & Daniel, B., 2007. Woven Surface and Form. AD Architextiles. Chichester: Wiley-Academy, pp. 82 – 89.

Von Foerster, H., 1982. Observing Systems. 2nd ed. Intersystems Publications

Von Foerster, H., 2003. ‘Understanding Understanding – Essays on Cybernetics and Cognition’, Springer, New York

Von Uexküll, J., 2009. An Introduction to Umwelt. In Hensel M., Hight C. & Menges A. eds. Space Reader: Heterogeneous Space in Architecture. Chichester: John Wiley and Sons.

Walter Grey, W., 1950. An Imitation of Life. Scientific American, 182, pp.42-45. Walter Grey, W., 1953. The Living brain. London: G. Duckworth.

Weisstein, E.W., 2013. Recursion. MathWorld: A Wolfram Web Resource, http://mathworld.wolfram. com/Recursion.html, accessed May 13th 2013.

Whitehead, H., 2003. Laws of Form. In: Kolarevic, B. ed. Architecture in the Digital Age: Design and Manufacturing. New York: Spon Press.

Wiener, N., 1961. Cybernetics: or the Control and Communication in the Animal and the Machine: Or Control and Communication in the Animal and the Machine. 2nd ed. Cambridge: MIT Press.

Williams, C., 2004. Design by Algorithm. In Leach, N., Turnbull, D. & Williams, C. eds. Digital Techtonics. Chichester: Wiley Academy,

Woodbury, R., 2010. Elements of Parametric Design. New York: Routledge

269


LIST OF ILLUSTRATIONS


LIST OF ILLUSTRATIONS

1| Introduction

[1] Generative drawing | by author

[2] Thesis structure | by author

3| Technology beyond Optimization

[1] Shipbuilding industry | Burtynsky, E., 2005. Shipyard #11, Qili Port, Zhejiang Province, China. In

Fishman, T, Kingwell, M. Burtynsky - China: The Photographs of Edward Burtynsky. Steidl Publishers, p.71 [2] Frank Gehry’s Walt Disney Concert Hall in production, Los Angeles, California | www.arcspace.com/ image-library/walt-disney-concert-hall/ accessed June 6th 2013

[3] Frei Otto, form found roof structure for Olympic Stadium and park, Munich Germany | Sandaker, B.N., Eggen, A.P. & Cruvellier, M.R., 2011. The Structural Basis of Architecture. London: Routledge

[4] Ashby’s Homeostat | www.vintagecomputer.net/electronic_brain.cfm, accessed June 6th 2013

[5] Algorithmic Network / Branching morphology| by author

[6] Model from the ‘Non-Linear Systems Dynamics’ workshop (led by Jenny Sabin, Peter Lloyd Jones, An-

drew Lucia, Erica Savig) at Smart Geometry 2010, as featured in the exhibition ‘Working Prototypes’ at DHUB, a contemporary design museum in Barcelona, Spain | photo by author

[7] Model from the ‘Non-Linear Systems Dynamics’ workshop | photo by author

4| A Manifold Digital

[1] Schumacher’s parametricism. Abu Dahbi Performing Arts Centre by Zaha Hadid Architects (2008) |

Betsky, A., 2009. Zaha Hadid Complete Works. London: Thames and Hudson

[2] Greg Lynn, Plan, Cardiff Bay Opera House Competition (1994) | Rappolt, M., 2008. Greg Lynn Form. New York: Rizzoli

[3] Greg Lynn, Model, Embryological House (1998-1999) | Rappolt, M., 2008. Greg Lynn Form. New York: Rizzoli

[4] Son-O-House, Nox/Lars Spuybroek (2000-04) | Spuybroek, L., 2004. Nox. London: Thames & Hudson [5] An analogue-computing model through which tectonics emerge from procedures of weaving and interlacing.

For Son-O-House, Nox/Lars Spuybroek (2000-04) | Spuybroek, L., 2004. Nox. London: Thames & Hudson [6] HygroScope: Meteorosensitive Morphology by Achim Menges and Steffen Reichert, 2012. | www. achimmenges.net, accessed June 6th 2013 [7-9] Dermoid | CITA

[10] OneMain Street by Mark Goulthorpe/dECOi Architects (2010) | www.decoi-architects.org, accessed June 6th 2013

[11] Hypnochamber, R&Sie(n) (2005) | Corbellini, G., 2009. Bioreboot: The Architecture of R&Sie[n]. New

York: Princeton Architectural Press

[12] ‘I’ve Heard About’-project, R&Sie(n) (2005) | Corbellini, G., 2009. Bioreboot – The Architecture of

R&Sie[n]. New York: Princeton Architectural Press

[13] ‘He Shot Me Down’-project, R&Sie(n) (2006-07) | Corbellini, G., 2009. Bioreboot – The Architecture of

R&Sie[n]. New York: Princeton Architectural Press

271


GENERATIVE PROCESSES IN ARCHITECTURAL DESIGN

[14, 15] Ali Rahim, Reebok Flagship Store, Shanghai, China (2004 - ) | Rahim, A., 2006. Catalytic Formations – Architecture and Digital Design. New York: Taylor and Francis

5| A Cybernetic Approach

[1] Lamella Flock, generative design system | by Martin Tamke & author

[2] Grey Walter’s tortoise | Pickering, A., 2010. The Cybernetic Brain. Chicago/London: The University of

Chicago Press

[3] The tortoise in action | Pickering, A., 2010. The Cybernetic Brain. Chicago/London: The University of Chicago Press

[4] The Fun Palace by Cedric Price | Mathews, S., 2007. From Agit-Prop to Free Space: The Architecture of Cedric Price. London: Black Dog Publishing,

[5] Gordon Pask on the cover of a publication from the Architectural Association | http://lebbeuswoods.wordpress.com/2011/09/28/heroes-of-a-revolution-gordon-pask/, accessed June 6th 2013

[6, 7] A three-dimensional cellular automata | by author

Experiment 1 | Lamella Flock

[1] ‘Broadway Boogie Woogie’ (1943) by Piet Mondrian | L’architettura. Cronache e storia., nr.5, 1991

[2] ‘Black Untitled’ (1948) by Willem de Kooning | www.metmuseum.org, accessed June 6th 2013

[3] Parametric Wood Construction (2007) by Martin Tamke (principal researcher) & Jacob Riiber | CITA

[4] The original Zollinger construction | Tamke, M., Riiber, J., & Jungjohann, H., 2010. Generated La-

mella. In: LIFE in: formation. On responsive information and ariations in architecture. ACADIA, pp. 340347

[5] The reciprocal pattern of the Zollinger construction | by Martin Tamke & author

[6] Restrictions imposed by the system and its production | by Martin Tamke & author

[7] Construction made by early prototype parametric system | by Martin Tamke & author

[8] Inspiration: Flocking behaviour | Wing-He. Murmuration. http://rc-4.net, accessed June 21st 2013 [9] The behavioural principles | by Martin Tamke & author

[10] Global behaviour: Negotiation of the agents | by Martin Tamke & author [11] Stabile configurations of the system | by Martin Tamke & author [12] Laser cut model produced from the Lamella Flock system | CITA

[13] The Lamella Flock demonstrator at ROM Gallery, Oslo 2010 | CITA

[14] Generated patterns for reciprocally supporting beam structures | by author

6| Representation and Agency

[1] Snowflakes | Prigogine, I., 1980. From Being to Becoming: Time and Complexity in the Physical Sciences.

New York: W.H.Freeman & Company

[2] Biological aggregation | Prigogine, I., 1980. From Being to Becoming: Time and Complexity in the Physical Sciences. New York: W.H.Freeman & Company

[3] Miscellaneous Mappings | Corner, J., 1999. The Agency of Mapping: Speculation, Critique and Invention. In: Cosgrove, D. Mappings. London: Reaktion Books

272


LIST OF ILLUSTRATIONS

[4] Iron filings around a magnet | http://commons.wikimedia.org, accessed June 7th 2013

[5] Barry Le Va, Bunker Coagulation (1995) | www.upenn.edu, accessed June 7th 2013

[6] Barry Le Va,, Continuous and Related Activities; Discontinued by the Act of Dropping (1967) | http:// whitney.org, accessed June 7th 2013

[7] Programmed drawing | by author

Experiment 2 | Porous Ascend [1] Conceptual model | by author

[2] The Penrose Pattern subjected to recursion | by author [3] Screenshots of the design tool in use | by author [4] System parameters | by author

[5] Extrusion of an element | by author [6] Process | by author

[7] Line drawing of the final demonstrator | by author [8] Unfolded elements | by author

[9-12] Final demonstrator | by author

7| Environment and Interface

[1] Ophrys Apifera, Orchid | Prof. Dr. Otto Wilhelm Thomé, 1885. Flora von Deutschland, Österreich und

der Schweiz. Gera, Germany

[2] Fluid simulation | by author

Experiment 3 | Scatterings

[1] ’Right After’ (1969) by Eva Hesse | Sussman, E. & Wasserman, F., eds., 2006. Eva Hesse: Sculpture.

New Haven: Yale University Press

[2] ‘Untitled (Rope Piece)’ (1970) by Eva Hesse | Sussman, E. & Wasserman, F., eds., 2006. Eva Hesse: Sculpture. New Haven: Yale University Press

[3] Examples of agent configurations | by author

[4] Properties evaluated when forming and negotiating connections| by author [5] Screenshots from the ‘agent space’ | by author

[6] Screenshots from the ‘simulation space’| by author [7] Components in moulds | by author

[8] Axonometric drawing of the component/mould construction | by author

[9] Final demonstrator as it appeared within the simulation space| by author [10] The process of making| by author

[11] Close-up of the demonstrator | by author [12] The final demonstrator | by author

[13] Close-up of the demonstrator | by author [14-15] Experimental probe| by author

273


Generative Processes in Architectural Design Jacob Riiber

The Royal Danish Academy of Fine Arts, Schools of Architecture, Design and Conservation School of Architecture

Generative Processes in Architectural Design PhD Thesis by Jacob Riiber


Issuu converts static files into: digital portfolios, online yearbooks, online catalogs, digital photo albums and more. Sign up and create your flipbook.