See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/235674111

NOX: Machining Architecture

Book · November 2004

CITATIONS READS
100 1,815

1 author:

Lars Spuybroek
Georgia Institute of Technology
29 PUBLICATIONS 333 CITATIONS

SEE PROFILE

All content following this page was uploaded by Lars Spuybroek on 28 October 2022.

The user has requested enhancement of the downloaded file.

part1.1-42.qxd 5/21/04 2:43 PM Page 1

machining
architecture
NOX Lars Spuybroek
part1.1-42.qxd 5/21/04 2:43 PM Page 2

14 Soft City 54 FOAM HOME

_ _
42 SoftSite 64 "beachness"
_ _
74 Tommy 114 OffTheRoad_5speed
_ _
158 D-tower 146 ParisBRAIN
Arjen Mulder _ _
332 The Object of Interactivity
174 Son-O-House, a house where sounds live 268 Students ANALOGUE COMPUTING
_ _
270 La Tana di Alice 272 Students CONFIGURATIONALISM
_ _

Lars Spuybroek
CHRONOLOGY
6 Machining Architecture
352 The Structure of Vagueness 378

Institute for Lightweight Structures

art research IL 9 – Pneus in Natue and Technics
IL 33 – Radiolaria
IL 35 – Pneus and Bone

Manuel DeLanda Frei Otto Gottfried Semper

370 Materiality: Anexact and Intense

architecture

Andrew Benjamin

342 The Surfacing of Walls

Detlef Mertins
exhibitions buildings
360 Bioconstructivisms

18 H2Oexpo 216 Soft Office

_ _
100 Flying Attic 46 blowout 234 Maison Folie
_ _ _
128 FACES 80 Goes goes 260 obliqueWTC
_ _ _
134 De Gothic Stijl
84 V2_Lab/V2_façade 280 FEDUROK
_ _ _
138 wetGRID 106 The Future Is Now 290 ECB
Brian Massumi _ _ _
322 Building Experience c
208 NOX FLURBS 198 Galerie der Forschung 304 Pompidou Two
_ _ _
part1.1-42.qxd 5/21/04 2:43 PM Page 4
4 5
Introduction
Introduction
The computer has reached a cultural stage, finally. The years that it
was used for dreaming of perfect shape grammars and design
automation (which only seemed capable of producing snowflake-
shaped churches and second rate Corbu’s) or worse, used for dream-
ing disembodied dreams of an architecture floating in cyberspace –
these years are over. Also finally, and especially important for the
architectural profession, have computers outgrown their servile func-
tion in the digital drawing room, where the real design was still done
far away from the machines, sketched by hand, guided by genius.
A computer is a steering device, more than anything else it is that.
According to the old rules of cybernetics this simply means two things:
direction and flexibility – and this book is about both. An architectural
design procedure that follows the rules of strategy and tactics needs
both the clarity of determinism and the fuzziness of variability. The
architecture of continuous variation is therefore not one of ‘free form,’
but of articulation and structure. Architecture is still an art of the line.
This book is a celebration of lines and of tectonics – complex lines and
complex tectonics I haste to say, since the computer generates and
manages complexity. During the second half of the twentieth century
science developed fully under the sign of complexity, otherwise known
as dynamic systems theory, which seems to have passed by architec-
ture fairly unnoticed. Architecture seemed to be caught in the local pol-
itics of language.
Machining Architecture aims to be three books in one: a monography,
a how-to book and a tractate. As a monography it shows the works and
projects as they were developed over the last ten years at my office
NOX. Art projects, research projects and architectural projects that
benefit from each other in a way where each of them seems to succeed
where the others can’t. And naturally they co-exist almost everywhere
in the book, the art projects inhabit the architecture and vice versa,
research and analysis in general are at the outset of every separate
project. As a how-to book Machining Architecture tries to be as clear
and transparent as possible on the design methodologies of ‘machin-
ing’: stepwise procedures of adding information into a system to gen-
erate form. In that sense it can partly be read as a cook book with tech-
niques and recipes that can be tested, developed or rejected. As a
tractate the book operates through essays, descriptions, notes and
short comments of myself and larger essays of ‘embedded authors’
interwoven with the projects. These thinkers were invited specifically
because of their own research, not to legitimize or criticize NOX’s work
but more to literally connect their own ideas to mine and make it click.
The book tries to develop a clear agenda. It states that an architecture
of complex, topological geometry can only be pursued through rigor-
ous means, and though its main theory is of vagueness, its practice is
of obsessive precision. It secretly dreams of a systems theory of archi-
tecture. Moreover, the topological turn means an innovation at a con-
ceptual level, as well as at a geometrical and a material level.
Generally innovation happens at one of these levels only, but not at all
three. This makes the type of work quite peculiar, for one there is not
enough building, for the other the theory is embarrassingly unacadem-
ic, for one the art makes it incomprehensible, while for the other the art
is just architecture without doors and toilets. It is not the aim of the book
to occupy a glorious middle position, though I am convinced that the
middle is more radical than anything.
L.S.
part1.1-42.qxd 5/21/04 2:43 PM
6 7
Machining Architecture
The title of this essay was used before in my
The Weight of the Image (Rotterdam: NAi
Publishers, 2001). That essay followed a
slightly similar path, but was fully based on
the notion of the diagram – a concept that
was highly influential in the 1990s but should
now be replaced because of its lack of real
precision and instrumentality.
1. Quoted from: Varela, Thompson and Rosch:
The Embodied Mind (London: MIT Press,
1997), p.174–175.
2. Maurice Merleau-Ponty, Phenomenology of
Perception (London and New York: Routledge,
2003), p.127 and p.165.
MACHINING ARCHITECTURE
Page 6
Lars Spuybroek
Action, Perception, Construction
In 1963, Richard Held and Alan Hein1 conducted a classic and rather merciless neu-
rological experiment in which two kittens were raised in the dark under very con-
trolled conditions. The main part of this conditioning existed in a physical carousel
where the kittens were restricted to move around only in a circle and were con-
nected to each other through a central pivot. One kitten was able to move freely
around a circular track, while the other was strapped in a suspended gondola,
which was pulled by the free cat that had its legs on the ground. As the young ani-
mals’ brain tissues developed, their actions and perceptions were integrated into
their own individual neurological systems. After a number of weeks, the kittens
were released from the carousel. The active cat moved and behaved normally,
while the passive cat stumbled and bumped into objects, and was afflicted with
agnosia – a condition of mental blindness brought on by neurological rather than
physiological causes. The free cat was able to link the act of walking to its own per-
ceptions, while for the other, action and vision were severed. It could not coordinate
its movements with what it saw because in its experience, action and perception
had never existed in the same continuum. The now famous and often quoted exper-
iment by Held and Hein proved that these two faculties were inseparable – per-
ception relied on action and action was only possible through perception.
Naturally the continuity of action and perception can only happen in a
body where “a perception is not followed by a movement, but where both form a
system which varies as a whole,” as Merleau-Ponty says2. That is, through the
Gestalt of a body-schema or body-image. This system is constantly fed by move-
ments, by actions, it coordinates and consolidates them – otherwise it would never
be a whole. Furthermore, the system needs to be plastic enough to incorporate new
movements. Somewhat later on in the text Merleau-Ponty gives the example of a
woman with a feather on her hat (it must be the early 1900s) who automatically
bends her head while passing through the doorway, and of the driver of a car who
automatically drives around obstacles and takes a curve without having to actual-
ly stop the car, take measurements and proceed driving after doing the calcula-
tions for the correct curve. The hat is you, the car is you. Through the body-image
we are not just capable of creating a system of interconnected actions that fit the
contours of our bodies but also to extend them into space. Then again, as the sto-
ries of Oliver Sacks make clear in such a terrible way, the opposite also applies
where we can shrink into bodies that are smaller; for example when the capacity
of using the legs is lost after being confined in plaster for a number of weeks. The
body-schema “is not fixed, as a mechanical, static neurology would suppose; the
body-image is dynamic and plastic – it must be remodelled, updated all the time,
and can reorganize itself radically with the contingencies of experience.” “It is not
something fixed a priori in the brain,” Sacks adds in an anti-Kantian swipe, “but a
process adapting itself all the time to experience.”3
3. Oliver Sacks, A Leg to Stand On (New York:
Touchstone, 1998), p.194.
It probably seems a bit odd to start the introductory text of an architecture book
with hard-core neurology and neurophilosophy but these examples and especially
the experiment with the cats always struck me as something very fundamental
being researched in relation to architecture. The fact is, what neurologically shares
a continuum, is architecturally still being considered as distinct. Are we architects
not trained to plan movement first and afterwards extrude it upward into the
image, that is, are we not trained to first draw the plan, the surface of action, and
then extrude it upward into the elevation, the surface of perception? Are we not
trained to treat floor and wall as discontinuous? Or, in addition, are we not trained
to treat walls, floors and columns as separated elements? Or should we, in paral-
lel to the body-schema, consider an architecture-schema as fundamentally plastic,
topological and continuous? Should we not consider this continuity between move-
ment and image as the “original curve”, as the Ur-curve that feeds action into per-
ception and perception into action? Then, when doing so, should we not also real-
ize that this curve is by its nature one of construction since it connects the horizon-
tal act with the vertical image?
But before we continue investigating further body-architecture relations
we should consider design itself.
The Need for Machining
“The organization of a machine (or system) does not specify the properties
of the components which realize the machine as a concrete system, it only
specifies the relations which these must generate to constitute the machine
or system as a unity. Therefore, the organization of a machine is independ-
ent of the properties of its components which can be any, and a given
machine can be realized in many different manners by many different
kinds of components. In other words, 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 in that space, and
have the properties which allow them to generate the relations which
define it.” – Humberto Maturana and Francisco Varela4
4. Humberto Maturana and Francisco Varela,
Autopoiesis and Cognition: The Realization of
The distinction between an organization and its structure, between a virtual organ- the Living (Dordrecht: D. Reidel, 1980), p.80
ization and an actual structure we should say, is very different from the Platonic
idea-and-form or Kantian scheme-and-reality oppositions. What these theories
have always failed to explain is how there can be real communication between the
two, that is, how there can be a physical relation between the two instead of a
metaphysical one. How can there be discrete organizations (like objects, which
part1.1-42.qxd 5/21/04 2:44 PM Page 8

8 9

5. Francisco Varela, Principles of Biological
Autonomy, (New York: Elsevier, 1979).
they are not); how can they be selected; how do they, when selected, become a
structure; and when it is a structure, why is it different from the others that fit the
same discrete set. These theories always needed an external body (God, archi-
tects) to activate the process, to enable the shift from the one side to the other,
because matter was considered passive and uncapable of passing through by
itself. But let us begin to consider things as being mobile themselves – and along
the way get rid of creationism in architecture. This notion involves the now well-
known concept of self-organization, where materials are active agents that seek
nothing but agency, that seek order, an order that is not transcendentally estab-
lished but emerges bottom-up. Though I immediately want to add here that this
emergence is always contained in a framework that is highly historical. The rela-
tion between organization and structure exists in all objects, either organic or inor-
ganic, either designed or grown – and for the purpose of the text it is important to
continue to mix them up. Clearly such an ontology (how do things come into exis-
tence? – shouldn’t we finally consider realism in architecture?) first needs to tackle
how an organization that is closed and singular can have multiple material struc-
tures that must be open or else they cannot vary. Or, as they say in biology, “a vari-
ation that is real of a type that is illusory” – we never see The Oak, we only see
oaks. Somehow, forces in the world are first capable of converging into an organi-
zational singularity and while passing through that point are then capable of
diverging into many different actual structures. Such an organization simply has to
be a topological structure or else it would not be able to change and create varia-
tions of itself. For instance, a typological structure is not capable of transformation
(only of deformation) since its components are fixed. The topological schema con-
centrates on the relations instead of the components. That is also why Maturana
and Varela are correct to classify them as ‘machines’ since their transformations
happen over a certain time period, they process transformations as a formative
procedure, as what Varela earlier had called in-formation5. Transformation is an
intensive process, and all changes are internal changes. Energy and time do not fly
about, they happen in material forms and structures, nowhere else (that is the con-
vergent part), but while it happens these forms transform and change (which is the
divergent part).
Let us now take a closer look at what this would mean for design. As already
indicated above the procedure has a convergent phase of selection and a diver-
gent phase of design.
• Contraction or convergence – a movement of virtualization, where infor-
mation is gathered, selected, graphed or mapped and then organized into
a virtual machine. A movement towards quality, order and organization.
• Expansion or divergence – a movement of actualization, where the organi-
zational diagram germinates and becomes formative. A movement towards
quantity, matter and structure.
Though this division clarifies what goes into the system and how it comes out, it is
still unclear about how the first stage is actually connected to the second. Though it
is evident at this point that it should be a continuous process and that it should be
phased, it still means we could take any diagram and drop it on any type of build-
ing to produce any form – which is simply the most horrendous thing I can imagine.
All the phases should be (empirical) machines in themselves. Machines connect
only to each other, as molecules, which means the phases in a process need to be
steps in a procedure. Finding rules. But when we use step-wise procedures doesn’t
it mean that the whole procedure is fixed in advance?6 Generally ‘procedure’ is 6. Bart Lootsma, “The Diagram Debate, or
the Schizoid Architect,” in: Archilab (cata-
used in a deterministic, mechanical manner, but it may be better to consider them
logue to the exhibition, Orleans, 2001), p.26.
as mechanical blocks that in themselves are quite straightforward and linear but
when linking up they can actually form non-linear strings. The blocks have complex
edges, multiple hinges and have many ways of connecting from one to the other
creating a complex chain of techniques. It is path dependent, like cooking, it works
with techniques and recipes, but it is quite uncertain what actually comes out (well,
when I am cooking at least).
Since this theory of form-generation is so highly dependent on material
processes let us rephrase the procedural techniques within a material framework
before we apply it more exactly onto architectural design strategies. Clearly the
above mentioned concepts of order, of how organizations transform into structures,
how they are extensive and metric first, then become topologized and then again
become form are deeply rooted in the ’intensive sciences’ as Manuel DeLanda7 7. Manuel DeLanda, Intensive Science and
coins them, like thermodynamics, topology and moreover dynamical systems the- Virtual Philosophy (New York: Continuum,
2002)
ory. Generally we (‘we’ architects) consider form as solid, but this is only a matter
of scale and often also of time or temperature, because material form has two
major properties: it is both flexible and stable. It is flexible enough to be moved out
of equilibrium and coherent enough to be moved back into equilibrium. That means
a theory of solids can hardly explain the spontaneous passage from organization to
structure. I believe it was Viollet-le-Duc who stated that architecture is an Art of
Crystals. Yet this would only support a theory of type, of archives and catalogues,
but not of self-generative form. In between we need liquid states, or at least more
viscous states that allow for reconfigurations. Now we can already be somewhat
more clear about the stages of design:
a. we need to select a system and create a configuration for the machine
based on this selection;
b. we need to mobilize the elements and relations in that system;
c. we need a phase of consolidation to finally have the system;
d. result into an architectural morphology.
From system to flexibility to rigidity to morphology. In the convergent initial phase
there is an acceleration, a topological mobilization that passes through a very nar-
row channel after which the process slows down and consolidates by taking a
divergent path towards a geometric form. As this still sounds quite obscure, we
need to clarify more and go through the stages one by one.
“We never think of transforming a helicoid into and ellipsoid, or a circle into
a frequency curve. So it is with the forms of animals. We cannot transform
an invertebrate into a vertebrate, nor a coelenterate into a worm, by any
simple and legitimate deformation. Nature proceeds from one type to
another … To seek for steppingstones across the gaps between is to seek in
vain, forever” – D’Arcy Wentworth Thompson8
8. D’Arcy Wentworth Thompson, On Growth
and Form (1917 and 1942). Quoted from:
So, all changes are small changes. And though they are intensive they have the Stephen Jay Gould, The Panda’s Thumb
power to either individuate or mutate (one never knows if one has just passed a (London: Penguin Books, reprinted 1990),
p.160.
bifurcation), or as Bergson would say, produce a difference in degree as well as a
difference in kind. Though a transformation can have a large effect it is always a

MACHINING ARCHITECTURE
part1.1-42.qxd
10
9. Manuel DeLanda, Deleuzian Ontology
(published on the internet). The idea of con-
vergence is strongly defended by Simon
Conway Morris, see his: The Crucible of
Creation (Oxford: Oxford University Press,
1998) and Life’s Solution (Cambridge:
Cambridge University Press, 2003).
convergence
consolidation
mobilization
morphology
system
machining
research
analysis
10. See my other text in this volume, “The
Structure of Vagueness,” p. 352
Column system for a tower generated by
wettened wool-threads, IL 35, p. 181.
MACHINING ARCHITECTURE
5/21/04 2:44 PM Page 10
relatively small step and the newness of the new can never be appreciated right
away. On the other hand, the quote from Thompson makes poignantly clear – more
than most of us would like to think – that type is relevant, not just in biology, but also
in architecture. Obviously when we set out to design a tower, for instance, we are
not going to start setting up a machine with a horizontal configuration. And though
we deeply need a topological technique to generate designs it will always be nec-
essary to topologize type (in our case of the tower topologize verticality), and not
just fly in a topological figure or system from wherever. I think there has been too
much emphasis on the divergent, proliferative capacities of intensive design tech-
niques and not enough on the necessary initial selective procedures. Because, if we
are going to mobilize elements and relations, in what system? Again, it was Manuel
DeLanda who gave us some clues for how to operate in these stages. First, he
advises to empirically look at real populations of buildings and second, he advises
us to rethink type as a ‘body plan’9. So, if one needs to design a tower (the same
one), look at all the towers, look at their diversity, analyse their differences, map
divergence them, organize them, look at their internal relations, look at their ‘body plan’. Body
plan as opposed to type means a much more abstract classification than pure
resemblance and identity.
In short, for self-generative design techniques we need empirical (since it
all happens within the real) research of already-existing forms, then we need to
construct body-plans out of this research through analysis, then these machines
need to be able to process information (or difference) through a mobilization of its
topologically connected components, then these need to be able to consolidate and
take on a form, first as a design and then as a real building.
Still continuing with the tower design, the phase of the machining procedure
building
(after research and analysis, which are convergent) would start with the actual
design
decision on the configuration of the system. In the book many examples are given,
but in this context it might be interesting to highlight the tower-generating machine
as it was developed at Frei Otto’s Institute for Lightweight Structures. Since it is not
the purpose of this specific text to fully investigate the complexities of these tech-
niques10 I will discuss them only briefly. What is relevant in this context is how the
machine is actually configured in relation to the notion of a body plan. The tower-
machine consists of two grids, one at the top and one at the bottom, at a certain
distance from each other. In between the grids are lines with a certain amount of
flexibility, or in this case where the lines are materialized as wool threads, a certain
amount of slack. After dipping the whole system under water and shaking it hori-
zontally (the mobilization phase) it is taken out of the water, and the threads imme-
diately self-organize into complex branching systems (the consolidation phase).
Now, we need to make clear that this is by no means The Method for tower-gener-
ation; it is just one of them, and we have made many variations (see p. 268). What
is important to observe however, is that the transfer from columns to wool threads
is not just a topologizing or flexibilization the structure but especially an abstraction
of a column system into a path system. In this case a path system for vertical loads.
Paths are capable of becoming single columns (when single threads become con-
crete or steel), thick columns (when multiple threads have merged), spatial
columns (where diagonals form nodes), diagonal columns (that generally come
later when bracing the verticals to the horizontal wind forces), mega-columns
(many diagonals form one large spatial tube), meta-columns (like center lines of
cores), etc. Paths are of a higher order, more abstract but as empirical as the
11
columns themselves. And because they are material (or programmed as material)
all movement is transferred in an intensive way, which means the movement dissi-
pates by transforming into structure.
Machining Architecture
In the book there are many techniques researched and developed that use the
mobilization of a system and consolidate it into form. The earlier techniques are
based on deformation principles, where a primitive like a sphere, a cylinder or sim-
ilar is bent and curved through step-wise processed movements. Though all these
movements are topological the technique is fundamentally of an indexical nature,
which means a ‘photographical’ registering by freeze-framing. Though there is a
considerable accumulation of time during the mobilization phase, the consolidation
phase is but a moment. This can only become structural when reread through
structural members afterwards that are then topologized as if they were part of the
mobilization itself. Later techniques, beginning especially with wetGRID which had
started after I had met Frei Otto in 1998, are different since they are more con-
structivist, where structural members join in and therefore focus more on the con-
solidating phase of the procedure. They are based on truly transformational princi-
ples, where often the consolidation is directly a self-supporting, a self-engineering
of the system. The system is not just shaken or deformed under the influence of
motion. It actually passes a critical threshold, a point of self-stopping, where it
transforms without being able to be reduced back to the original state. The often
symmetrical state of the initial set-up is fully broken into new complex geometries.
Later projects are more of what I call a configurational nature where the move-
ment-phase is replaced by a limited set of variable figures (S-curve, O-curve, L-
curve, U-curve, etc.) and the consolidation-phase is replaced by a matching of
these figures first in pairs (translation, rotation and reflection) then in larger con-
figurations. While the indexical technique allowed us to be very exact during the
mobilization phase, and the constructivist technique gave us more precision for the
consolidatory stage (and by nature being more emergent), the configurational
method combines the best of both. While being based on the same notion of bro-
ken symmetry the figure-configuration method allows for a more precise calibra-
tion of formal, structural and programmatic information. But before we will pro-
ceed with matters of design, i.e. how machining relates to action (program), per-
ception (form) and construction (structure) we should investigate the fourth phase
of the machining procedure, morphologies, being highly dependent on the above
described techniques.
“How did we ever get the little bones in our inner ear, which directly origi-
nate from the jawbone in reptiles? That is unimaginable. The creationists
would say that this transformation is impossible because during the trans-
fer of these bones from the hinging position in reptiles to the inner ear of
mammals the lower jaw would be hanging loose. This is of course not the
case. The transformative shapes have a double maxillary joint, so the func-
tionality is kept when one of them moves to the inner ear position. There is
always this huge redundancy. This leads of course to an organic machinery
that is everything but optimized, seen from the traditional principles of
human design” – Stephen Jay Gould 11
11. Stephen Jay Gould, quoted from an inter-
view in: Een schitterend ongeluk [A splendid
accident] original Dutch, Amsterdam:
Contact, 1993), p. 384
part1.1-42.qxd 5/21/04 2:45 PM
12
12. Peter J. Bentley ed., Evolutionary Design
by Computers, (San Francisco: Morgan
Kaufman, 1999).
13. “The Structure of Vagueness,” p. 358.
14. Brian Massumi, “Building Experience,” in
this volume, p. 328.
15. It is inter-dimensional – dimensions are
nothing but the organization of movement.
The procedure of system (lines), techniques
(line to surface) and morphologies (surface to
volume) should be viewed as a building of
movement upon movement.
deep
surface
blister
filo
sponge
16. Actually we should consider three levels
of machining: a. at the passage of organiza-
tion to structure, b. at the passage of design
to building, and c. at the passage of building
to experience. The middle level, that of com-
puter controlled machining techniques, has to
be left aside here. The same applies to the
differentiation between plastic and electronic
interactivity on the third level. However, all of
these items are intensely discussed through-
out the book.
MACHINING ARCHITECTURE
Page 12
Everything but optimized! This is the morphology of the provisional not the optimal.
Often generative techniques are being proposed under a sign of efficiency and
optimization.12 However, since the generative relies wholly on the topological, and
since the topology is real and fully materialized, one also gets the less determined
in-betweens.13 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 tran-
sitive geometry.14
The systems discussed in the book are generally line-systems, the techniques are
almost always line-to-surface techniques (though there are exceptions), and
almost all the morphologies are surface-to-volume geometries. The procedure as
a whole follows the classic drawing-to-building transfer. It is only since the dimen-
sions are not given beforehand and emerge afterwards that they turn out as being
continuous instead of being discrete. In other words, with a transitive geometry the
dimensions of building are not mechanically added up but organically synthe-
sized.15 In an architectural framework this generally means that the transformation
takes place in the more generic, simpler states of specific body plans (plane
façade, plane floor, row-rooms, row-houses, floor-roof, stacked floors, etc.) to
progress into more complex structures. In the book there are only a specific num-
ber of morphologies that come to the fore:
a. Deep Surfaces. Surfaces either on a flat or curved plane that are deformed
generally perpendicular to the direction of that surface. They can be tubes
deforming (H2Oexpo, BlowOut, OfftheRoad, Tommy) or flat planes deform-
ing (V2_Lab, Maison Folie Lille). A subset of the first category would be
porous surfaces where the deformations are structurally transformed
because of local (Fedurok) or global (SoftSite, ParisBRAIN) tearings.
b. Blisters. Single flat surfaces that are locally transformed to become vol-
umes. It is both deformational like the bulging effects of the first group as it
is transformational (De Gothic Stijl, wetGRID, GDF, D-tower, La Tana di Alice,
ECB).
c. Filo. Either double layered filo (Soft Office, Son-O-House) or multi-layered
filo (Pompidou Two). The first one is a double-layered blister-system where
the pockets lock into each other in such a way they start sharing curvature.
This happens even more in the multi-layered system where stacks are diag-
onally connected through packed structures of sponge-like morphology.
d. Sponges. Pure multi-oriented morphology either vertically stretched (o-
bliqueWTC) or unstretched (FoamHome, The Future is Now, “beachness”).
Experiencing the Machine
In the end, these machines are for bodies to experience them. How are they relat-
ed? Before we looked at how what goes into the machine is related to what comes
out. Now we should look at how what comes out of the machine is related to what
goes in.16 We first return to the statement from Oliver Sacks on the plasticity of the
body-schema: “It is not something fixed a priori in the brain, but a process adapt-
ing itself all the time to experience.” So, what is an experience (in the now)? And,
13
how does that relate to experience (from the past)? In architecture we are
obsessed by habits, and rightly so, but often we mistake them for mechanical acts,
for usage. Habits should not be understood as the sculpting of a passive schema
that only archives its actions to enable only their exact repetition. We are not
machines, we just want to be machines. The schema consists of rhythms and peri-
odic patterns, and it is exactly these which enable variability or change. In
Phenomenology of Perception Maurice Merleau-Ponty introduces the concept of
abstract movement, a movement-tension that is always present in the body. It is the
movement that is available in the body, a “background tension” he calls it, follow-
ing the Gestaltists from the 1920s. In fact this movement only becomes available to
the body through the numerous actions performed in everyday life – movement is
made up of movement, and the abstract and the real feed back and forth continu-
ously. Every act springs from this background tension, a real, actual movement that
“releases itself from neurological anonymity.”17 In short, it is a formidable critique 17. Maurice Merleau-Ponty, Phenomenology
of Perception (London and New York:
of the architectural program as the mechanistic layout of human behaviour within Routledge, 2003), p.127
a built system purely viewed as tasks, routines and habits. With the idea of rhyth-
micity – “an internal music” as Sacks calls it – an act is never completely certain, it
always differs from itself, and is always ready to shift into another act, or even to
slide into a ‘free’ act. When every act, however intentional, is also orientated side-
ways a lot of in-between program could unfold, both undetermined and unpro-
grammed program. Could unfold, since that depends on the architecture.
Now, this does not mean there is no ‘program’, of course there is – how else
create in-between program? – like there is habit and routine, or to put it even
stronger, habit and routine are at the outset of all design. It is only that we need to
consider them as potentially flexible, with a flexibility that is limited to different
degrees. Therefore it is very important to set the range of the variations, the
degrees of freedom beforehand, since they need to be productive in a selected
field only. Meaning, not every wall-floor connection should be that of a curve. But
there should be no mistake in comprehending that the actual feeding of movement
(during the ‘mobilization phase’) is that of human action. Secondly, this movement
can only be reciprocally fed back into human action when abstracted into structure
(during the ‘consolidation phase’), thus, abstraction of movement is building, noth-
ing else. There is a fundamental constructivism. Every engineer understands this
and every architect understands this – but they often do not understand it the same
way. This ‘movement’ is simultaneously of a single body, of groups, of multiple
groupings, and of forces and loads. There is an engulfment of schema all happen-
ing at the same time. Bodily posture (which is that which coordinates vision and
action) is fundamentally a constructive act by itself, both in the concrete connection
of feet and eyes as in the abstract rebuilding of the schema. But posture in all its
variations should be viewed in relation to social groupings, and is as molecular and
material, varying between crystallized states and liquid states, states of configur-
ing and reconfiguring. Then, what else can a building be? What else can it be than
a permanent becoming-built of the architecture-schema, and what else can the
architecture be than a permanent interaction with body-schema and group-
schema? All schema are continuously rubbing over each other, during the design as
well as during the experience when finally built, they feed movement and structure
into each other, channeled by perception – they constantly feed action and vision
into each other, channeled by structure.
part1.1-42.qxd 5/21/04 2:45 PM Page 14

14

Soft City Soft City

video (7’03”), broadcast on national television 28 March 1993
and exhibited at several international festivals,
VPRO television, 1992 –1993

Generally movement is considered to be in betw een the

images. In f ilm or in arc hitecture we are al ways con-
fronted with series of imag es that are mo ved or w here
we have to pass by our selves. By mor phing we could
finally have the mo vement in the imag e itself. In the
video movement of v ehicles is inter nalized as c hange
and transformation, and becomes related to growth and
living form – in the end the g enetic experiment arrives
at a city that suddenly starts to mutate. In a nutshell our
program for architecture of the ten years after 1993.
part1.1-42.qxd 5/21/04 2:45 PM Page 16
part1.1-42.qxd 5/21/04 2:45 PM Page 18

18

H20expo H20expo
water-experience pavilion and interactive installation for ‘WaterLand’ in a
private/public partnership with the Dutch Ministry of Transport,
Public Works and Water Management.
Neeltje Jans, The Netherlands, 1993–1997

The water pavilion is a v ery special buil ding for many

reasons. It was the first fully topological structure where
curvature is not onl y in the r oof and w alls, but also as
much in the fl oors: no section is hor izontal. It w as also
the first fully interactive environment where visitors can
transform light and sound in the interior through a wide
range of sensor s.

The water pavilion, H20expo, ia a permanent structure which does not ‘contain’
an exhibition in the classical sense, as in a museum where moving and seeing
are separated. Instead, the images and sounds that emerge depend on the
activities of the visitors. And the activities of the visitors depend on the images
and sounds that are always changing. The interactivity is a much larger concept Generally people only know double curvature from their car or their ca t,
here than the general technological understanding of the word. Interactivity the reason w hy this arc hitecture is often mix ed up wi th the sha pes of
here does not just mean that the building is ‘wired up’ into an environment of either vehicles or animals . The first is the ol d motif of streamlinin g, the
transforming atmospheres through electronic interventions, but moreover an second of biomor phism. Both have become extinct because they operate
architecture of transformation itself. only on the outside of objects . With streamlining all movement is pr o-
The geometry is generated by an iterative procedure of transforma-
jected on the sur face and does not result in an y internal reconfiguration
tions. It starts with a simple tube made up of ellipses. Then these ellipses are
and the second is pure mimi cry that also lacks a str uctural understand-
rescaled according to the needs of the program, and then deformed again
ing of form. A building is not a car wi thout wheels or a cat without legs.
according to influences of the site, like the main wind direction, position of the
sand dunes and orientation on incoming visitors. At the entrance of the building
it begins with a small ellipse on its vertical axis and it ends, some 60 meters fur-
ther, with a much larger ellipse on its horizontal axis – in between the building
twists and turns. Since the sections are mostly continuous it means that the floor
blends into the wall, and the wall into the ceiling and that no floor is horizontal. The first set of operations: elliptical tube, scaling of tube according to program, twisting
On every moment the visitor has to rely on his or her own motor system to bal- according to exterior forces, insertion of ground level, deformation of ground surface.
ance (since there are no windows looking outside onto the horizon and inform-
ing the body what is level): he or she has to act like water to pass through the
‘building.’
The first part of the building, which starts with a three-dimensional
door, is constantly flooded with water in different ways. After entering, one
part1.1-42.qxd 5/21/04 2:45 PM Page 20

20 21

comes to the ‘glacier-tunnel,’ which is completely frozen, with its melt water
spilling over the floor. A bit further into the building there are ‘springs’ spraying
mist and water, a ‘rain bowl’ with stroboscopically illuminated rain and the
‘well.’ The ‘well’ contains 120,000 liters of water and has its own program of
projections and light. As an object it puts everything else out of balance and is
one of the main generative forces on the structure. The ‘well’ becomes another
kind of horizon, an inner horizon, not horizontal but vertical, the axis of vertigo,
of falling since imbalance is the main sensation of one’s passage through the
building.
The interactivity of the architecture is continued by the interactive
installation of projections, light and sound. In the building specially designed
sensors are connected to three different interactive systems that operate
together. A system of real-time generated animations connected to a number of
LCD projectors, a system of some 200 blue lamps placed on the cable duct, and
a sound system which can interactively be manipulated and changed.
The sensors are differentiated for crowds (light sensors), groups
(pulling sensors), and individuals (touch sensors). Every group of sensors is
connected to a projector that shows a simple wire frame grid which translates
every action of a visitor into real-time movement of (virtual) water.
The light sensors are connected to the ‘wave.’ Every time one walks
through an invisible beam of infrared light the projected wire frames starts
undulating. With four of these sensors, the visitors can create any kind of inter-
ference of these waves. The touch sensors create ‘ripples’ in the wire frames
and pulling sensors are related to the ‘blob,’ a wire frame projection of a
sphere, which can be manipulated like a drop of water in zero gravity: it
bounces and stretches while pulling the sensors. These projections are scattered
around the building, but the effects of the sensors are also combined on a spine
of light, a 60 meter long row of blue lamps. First, without activating any of the
sensors, a wave of light is produced by these lamps, a ‘pulse’ that accelerates
every time a light sensor is activated. So, the more people that are in the build-
ing (and pass the invisible light sensors) the speedier the light undulates
through the building. The touch sensors likewise create pulses on the light curve:
a sudden high level of light splitting and fading in two directions. All these dif-
ferent light waves interfere with each other on the line. Finally, the pulling sen-
sors (for the ‘blob’ projection) are able to slow all this down. When people exert
maximum force on the sensors the light curve will be frozen in its last position.
Next to the light and the real-time projections, is the sound system. All sensors
are connected to a CD-Rom with sound samples that can be deformed, bent,
and/or stretched. But also the sound itself can, for instance, be electronically
pulled out of the ‘well,’ or pulled towards one self on the spine of light, which is
also the basis of the sound system.
part1.1-42.qxd 5/21/04 2:46 PM Page 22

22 23

The final operation in the design procedure is

the renegotiating of the ground level within the
building itself. The sections also show the circle
segments constituting each of the trusses. Each
circle segment fits exactly with the adjacent
one, making it a smooth curve. When designing
the water pavilion we worked in AutoCAD 11, a
program that could define ellipses only by cir- If one ‘closes’ two curves with straight lines (which is called a ‘ruled surface’) these lines seem quickly trans-
cle segments. Completely wrong, but fortunate- lated into a secondary structure. But since the tangent on one end of the beam is not the same as the tan-
ly so, because the machines rolling the HE-A gent on the other end one cannot use a normal rigid beam. Our steel contractor explained to us that the
steel profiles could only work with circles. beams he used were so cheap (and weak in the lateral direction) that they would simply torque while bolt-
ing them to the primary, curved trusses. He was completely right, all the 280 secondary beams are
torquing, sometimes up to 50 degrees.
part1.1-42.qxd 5/21/04 2:46 PM Page 24

24 25
part1.1-42.qxd 5/21/04 2:46 PM Page 26

26 27
The ripples are calculated by so called ‘reality engines,’ graphic cards that allow for
incredibly fast image generation. When the sensor is pushed with a certain force, the
computer immediately ‘drops’ a virtual stone of corresponding weight in virtual
water. The projector is set up in such a way that the center of the ripples created by
the virtual stone coincide exactly with the position of the sensor, i.e. your foot or hand.
We actually programmed the force/weight ratio in a non-linear way, so that children
get as much effect as adults.

Electrical diagram of connected comput- touch sensor

ers, projectors, loud speakers and sensors.
infrared sensor

The idea behind the electronic interactivity

pulling sensor
is that the system would both respond
locally and as a whole. Also that it would loud speaker

respond to passive visitors as much as to

computer
very active ones. Three interactive
processes operate simultaneously: wire- projector

frame projections, light movement over the

central curve and sound. In total there are wire frame

six wire-frame projections: four ripples,

light curve
one blob and one wave. The wave is acti-
vated by visitors simply passing the invisi-
ble infrared sensors, the ripples are acti-
vated by groups of touch sensors and the
blob is a game between four players that
each have their own pulling sensor. All
these manual operations have an immedi-
ate, real-time effect on the wire frames:
waves in line- or in circle-patterns are
transferred through the mesh. It is not a
film: the ripples are real-time calculated
waves dependent on the forces the visitor
uses to manipulate the sensor.
part1.1-42.qxd 5/21/04 2:46 PM Page 28
part1.1-42.qxd 5/21/04 2:46 PM Page 30

All the interactive effects relate to one another. From top left to bottom right: the blue
light moves by itself over the light curve hanging at the ceiling of the water pavilion.
Then, when some visitors pass the invisible sensors they create both an undulation of
the wire mesh projected on the floor under their feet and a speeding up of the light
pulse. In the middle column one can see an extra manipulation of two sensors that cre-
ate local ripples in the projections, and simultaneously create more water effects on the
light curve. On the right column extra sensor effects are introduced illustrating a typi-
cal day, when many visitors operate the sensors.
part1.1-42.qxd 5/21/04 2:46 PM Page 32
part1.1-42.qxd 5/21/04 2:47 PM Page 34

34

The first third of the fresh-water pavilion concerns real water: small springs, a jumping
water jet, and a rain bowl that sprays rain that when stroboscopically lit seems to fall
upward. All the water is spilled out over the floor making the visitors jump away in
excitement and laughter.

The diagram shows clearly that this building hardly has a program in the classic architectural sense. The
visitors are like water molecules, sometimes moving as individuals, sometimes moving in small groups, in
excited packs or passive rows, and sometimes moving in large crowds.
part1.1-42.qxd 5/21/04 2:47 PM Page 36

37

The middle area of the fresh-water pavilion is dominated by the Well, a

huge structure that contains 120,000 liters of water. A projection of a drop
of water falling is being projected in slow motion at the bottom of the Well.
In the interactive soundscape it produces a deep sound of extremely low
frequency. From the Well, the path splits to where one direction leads down-
ward into a dark wet cellar where all natural water is gathered to be filtered
and the other direction where the full interactive electronic part starts with
the wave projection.

During the design I became very interested in neurophilos-

ophy. Neurology seems to be the only branch of science that
by its nature coincides with philosophy. Bergson knew that,
as did Mer leau-Ponty. With H20expo it became clear to me
that questions of posture , of perception and acti vity are
architectural questions. We then called it ‘motor geometry,’
the abstract movement in the building, with its transforma-
tive geometry, relating directly to real mo vement of the
body.

p 97
p 139
p 227
part1.1-42.qxd 5/21/04 2:48 PM Page 38
38
When all architectural elements are connected through geometri-
cal continuity many social effects emerge that are totally unex-
pected. For instance, an elderly man stood in front of the larger
bumps, paused a moment, and suddenly ran up the slope. This
slope was meant for projection, not for walking. Clearly the
abstract movement of topology intensifies sensations that instigate
39
There is a per sistent misunderstanding of the
architecture-movement relationship. Through
topological vagueness architecture can acquire a
language of mo vement, i.e. ‘splitting,’ ‘merging,’
‘bending,’ ‘twisting,’ etc., tha t enables the archi-
tecture to move without the actual mo ving of the
building.
decisions in the body. There is more tension. It creates a larger
potential of movement without prescribing specific actions. Events
are not functions or mechanical actions anymore, they now emerge
from the interaction of a more undetermined architecture and the
body.
The topological vagueness is not general, not unarticulated free
form. Already in H20expo there are many internal edges to the land-
scape. Like in many later projects the edges are determined objects
taken up in a larger fields of less determined morphology.
part1.1-42.qxd 5/21/04 2:48 PM Page 40

View publication stats