The Frame And The Generic Space, A New Way Of Looking To

Flexibility
Dr. ir. Bernard Leupen, Associate Professor
Department of Architecture, Delft University of Technology
Berlageweg 1, 2628 CR Delft, Netherlands. B.A.J.Leupen@bk.tudelft.nl

Abstract

Houses have an average life span of about a hundred years, whereas households and
habitats can change radically and repeatedly during that time. Consequently house
designers are faced with the task of giving form to a shelter for dwelling for a period
during which the composition of the household and the associated spatial rituals will go
through major changes.

Taking not the changeable but the permanent as a departure-point opens up new
perspectives. The permanent, or durable component of the house, constitutes the frame
within which change can take place. This frame defines the space for change. The frame
itself is specific and has qualities that determine the architecture for a long period of time.
The space inside the frame is general, its use unspecified; this space I have called generic
space.

In this sense the frame frees other parts of a building. Take, for example, the loadbearing
column. It relieves the wall from acting in a loadbearing capacity, it frees the wall. A
notion essential to the frame's functioning is that of disconnection. The column can free
the wall by virtue of the fact that wall and column are not inextricably linked.

A building can be separated up into a number of layers that together defines the building
as a whole. Accordingly, the building can be regarded as a composition assembled from
these layers. Each layer is distinguished from the others by the special role it fulfils. In
the frame concept it is assumed that every layer may in principle serve as a frame. Basing
my information on texts by Laugier, Semper, Loos, Duffy and Brand, I have made a
distinction between the following five layers: Main load bearing structure, Skin, Scenery,
Servant elements, Access. In principle I distinguish three categories of changeability: the
alterable, the extendable and the polyvalent. These three forms of changeability can be
linked with three types of generic space.

To explore my concept, I present an overview of every imaginable combination of layers.

This catalogue of frames is then divided among four distinct series of combinations. The
basic combinations and the combination series, constitute the tools for designing houses
that proceed from the frame concept. It was the intention of this study to develop the
frame concept and the body of concepts attendant on it. Building upon its predecessors, I
developed a stimulating resource for anyone involved in designing houses that are able to
accommodate change. The potentials and limitations of the frame concept can be further
explored as designing proceeds.

Keywords: Frame, Polyvalence, Layers, Generic space, changeability

1
Introduction

This study1 is grounded in the fact that houses have an average life span of about a
hundred years, whereas households and habitats can change radically and repeatedly
during that time. Consequently house designers are faced with the task of giving form to
a shelter for dwelling for a period during which the composition of the household and the
associated spatial rituals will go through major changes.

Flexibility and changeability are key words in the approach of design for the
unpredictable. When the architects of the modern movement at the beginning of the 20e
century where facing the problems of mass housing flexible floorplan became a topic.
Specially the concept of the minimal dwelling and efficient use of the space generated
new solutions with sliding doors and folding beds providing the change from day to night
floorplans. In the Netherlands Mart Stam and Johannes van den Broek designed houses
(see figure 1) based on these ideas (Ottenhof 1981 (1936)). Beside of the day and night
floorplan some architects at that time developed concepts for the free plan like the so cold
plan-libre of Le Corbusier and later the support concept of Habraken.
Figure 1. Floorplan housing designed by J. van den Broek

Taking not the changeable but the permanent as a departure-point opens up new
perspectives. The permanent, or durable component of the house, constitutes the frame
within which change can take place. This frame defines the space for change. The frame
itself is specific and has qualities that determine the architecture for a long period of time.
The space inside the frame is general, its use unspecified; this space I have called generic
space.

The notion of frame is informed by the book Earth Moves by the French architect and
philosopher Bernard Cache (Cache 1995). One of Cache's assertions is that architecture is
the art of the frame. He distinguishes three functions that the frame performs: it separates,
selects and rarefies. In the present study I propose that the frame has a fourth function: it
frees. Take, for example, the loadbearing column. It relieves the wall from acting in a
loadbearing capacity, it frees the wall. The non-loadbearing wall can then be moved

1
This text is based on my PhD Thesis (Leupen 2002)
2
freely. A notion essential to the frame's functioning is that of disconnection. The column
can free the wall by virtue of the fact that wall and column are not inextricably linked, in
other words they can be disconnected.

Of what the frame might consist of? A building can be separated up into a number of
layers that together define the building as a whole. Accordingly, the building can be
regarded as a composition assembled from these layers. Each layer is distinguished from
the others by the special role it fulfils. In the frame concept it is assumed that every layer
may in principle serve as a frame. Basing my information on texts by Laugier (1977
(1753)), Semper (1851), Loos (1962 (1898)), Duffy (1990) and Brand (1994), I have
made a distinction between the following five layers:

Main loadbearing structure (columns, beams, loadbearing walls, trusses and structural floors). The
load bearing structure transmits the loads to the ground.
Skin (facade, base and roof). The skin separates inside and outside and at the same time represents
the building externally.
Scenery (cladding, internal doors and walls, finish of floors, walls and ceilings). This scenery
defines the space including its visual and tactile qualities.
Servant elements (pipes and cables, appliances and special amenities). The servant elements
regulate the supply and discharge of water, energy and air and also include the appliances
necessary to them and the spaces primed to accept these.
Access (stairs, corridors, lifts, galleries). This layer takes care of the accessibility of the spaces
and/or the individual homes.

Aided by various sources including Priemus (1968) and Elsdonk (1990), I have done
research into the forms the changeable can take. In principle I distinguish three categories
of changeability: the alterable, the extendable and the polyvalent. These three forms of
changeability can be linked with three types of generic space. Should the generic space
contain a layer that can be changed then we may describe it as alterable. Should the
generic space not be bordered on all sides then it is a question of extendibility. Should the
generic space contain no other layers while the generic space invites different uses
through its form and dimensions, then we have polyvalence; the generic space is then a
polyvalent space. My research is based in the first instance on knowledge registered in
designs and realised buildings. To be able to 'read' this knowledge from them it is
necessary to analyse these plans. For this we need an analytical tool that is focused on the
knowledge in question. The subdivision into layers as I proposed, is a sound tool for such
analysis.

So as to better get at the properties of layers and frame, I have researched the five layers
in their development into independent layers. The developments undergone by
loadbearing structure, skin and scenery are so closely bound together that the
development of these layers has been described as a whole. During the process, these
layers become independent and regroup to form new coalitions and then become
independent again. If we can discern in the primitive hut and the timber-framed house
that grew out of it two distinguishable layers (loadbearing structure and skin), in the 18th-
century house they were joined by a third layer, scenery. With the refining of applied
materials and the shift in architectural ideas about exposing the structure, this layer would
partially fuse again with the loadbearing structure and the skin.

3
A new generation of buildings came into play with the emergence of the iron skeleton. In
the first generation of these buildings the facade acts in a load-bearing capacity. Skin and
loadbearing structure then together constitute the frame. This I designate with the term
integrated frame. This notion denotes frames consisting of two or more integrated layers.
Such buildings can house different programmes without the need for radical
constructional measures; the generic space is, then, a polyvalent space. During the course
of the 19th century the skin became increasingly distinct from the other layers, a
development that in a structural sense reached a provisional end with the application of
terracotta elements by members of the Chicago School (see figure 2). However, fire
regulations prevented the skin from achieving complete independence from the steel
skeleton.
Figure 2. Burnham & Co. Reliance building, Chicago. Fragment of the floorplan and section.

New possibilities presented themselves with the arrival of non-flammable reinforced

concrete. Le Corbusier was one of the first to explore the architectural qualities of the
concrete skeleton. His Dom-ino study primarily concerned a division between carcass
and finish, in his later villa designs the skeleton and in particular the column became
more and more set apart and articulated as an architectural element, while the scenery
also gained its autonomy.

Duiker and Bijvoet, in their design for Zonnestraal aftercare colony, succeeded in
developing and articulating the concrete skeleton in all its facets into an architectural
object (see figure 3). Ostensibly Van Eyck is embroidering further on this development
with his Orphanage but closer consideration reveals a shift occurring here between the
assemblage of elements that structurally define the frame and the assemblage of elements
that represent it (see figure 4). The architrave used here represents the loadbearing
structure but is itself part of the skin or the scenery, depending on its position. It can also
happen that a number of layers or portions of layers together form a frame. This is true of
the Centraal Beheer office building. In situations like these I have introduced the notion
of the combined frame. Since the 1970s the increasing demand for insulation has
necessitated excluding even concrete loadbearing structures from view. Here too the

4
facade can represent the loadbearing structure. The emancipated glass facade wrapped
around the loadbearing structure gives the imagination free rein, conjuring up images that
might refer to what is inside the building but might well have other meanings.

Figure 3. Duiker and Bijvoet, Zonnestraal aftercare colony

The frame's articulation and the freedom that the frame creates is an important aspect of
the frame concept. In principle it involves a more general issue, namely the relationship
between flexibility and expressive architecture. Mies van der Rohe felt that a flexible
building demanded a high-powered architecture (Spaeth 1985) p. 117. Van Eyck by
contrast was of the very opposite opinion. The irony is that the Orphanage proves that
Mies was right. The Villa Savoye, however, shows that Mies van der Rohe's postulation
is not automatically reversible, for a forcefully articulated building does not necessarily
make it a flexible one and, by extension, a frame. In spite of the column structure this
building is flexible at all.
Figure 4. Van Eyck Orphanage, Amsterdam. Decomposition in layers

5
After a description of the development of loadbearing structure, skin and scenery, it is the
turn of the servant elements and access. The servant elements consist of piping and
ducting, the associated necessary appliances and the spaces primed to receive these.
Banham (1969) has pointed out that the development of servant elements has taken place
largely outside the architectural debate. Only during the course of the 20th century do we
see a tendency to express this layer in the architecture. In some cases the organization and
form of the servant elements go on to generate freedom for the other layers and so the
servant elements work as a frame. Accommodating this layer in a zone or core produces
spaces that are devoid of pipes, cables and appliances and therefore generic. A space
uncluttered with service elements gives great freedom for scenery and use.

From the Middle Ages onwards access to a house evolved from a single corridor or stair
into a system in its own right, an independent layer in the architecture. The prime driving
forces for separating living space and circulation space inside the house were privacy and
a sense of embarrassment (Elias 1990 (1939)). The desire for higher densities and the
concomitant stacking of dwellings led to ingenious systems of stairs, corridors, landings
and galleries. The access, originally interwoven with the floor plan of the house,
gradually became disconnected from it. At the same time access entered into a new,
structural alliance. Stairs and galleries were then made of concrete and merged with the
loadbearing structure.

Next, the Modern Movement provided the architectural means to give the access a
countenance of its own. Porch stair and gallery were articulated individually in the desire
to represent the programme in the facade, though the corridor was still withheld from
view. From the examples I analysed it transpires that the access can only be elevated to a
layer of importance if given an extra function. This function may be representative (the
foremost stairwell of the Haussmann block), or the access may be designed as a street
(Spangen), balcony (Koekoekstraat) or front garden (Golden Lane). In this respect the
gallery and porch have more potential than the corridor.

Now I enter into an aspect of the defining of layers. I suggest that every layer can imply
one or more further layers. Every facade has its own loadbearing structure, a servant
space can have its own scenery, and so on. The existence of these subsystems or
sublayers can be seen as fractalising of layers, a notion culled from chaos theory. In
discussing the five layers within the wide field of architecture the question arises of what
has been developed in the way of changeability in the province of housing. I have chosen
to use the four characteristics that particularise housing design as the leitmotif. These four
fields I have designated with the terms compartmentalisation, access, service system and
spatial division.

In the cases Obus (see figure 5), Unité, Urban Megastructure we can identify large
structures that deal with access to individual units as well as with supporting them as a
whole. Habraken developed the concept of supports (Habraken 1961), relating to large
structures to be developed by the community within which each unit is built to the
occupant's specifications. With a switch in the building market from high-rise to low-rise,
attention shifted from the large combined support and access structures to
compartmentalization and changeability within the compartment. The support concept

6
thus ceded to the carcass concept. Such as the fully stripped-back loadbearing structure
(Haaksma's Casco project) to the polyvalent carcass (Hertzberger's Diagoon houses).
A number of experiments done with the changeability of dwellings proceed from the
problem of services, the piping and ducting. In this field three approaches are possible:
accommodating the services in raised floors, in columns and in facades. In some of these
experiments the solution was sought in accommodating the pipes and cables in some way
in the frame, the very frame that now had to bring about their freedom. It is this
ambivalence that can result in additional investments. In dividing up the space - the
fourth problem area - scenery plays a key role. It is conceivable that this layer too can
enable changeability in the house and its use. This could happen with the aid of scenery
with many movable parts. It is open to question, however, whether such 'active' scenery is
the only way to design a house than can be used in different ways, i.e. is polyvalent.

Figure 5. Le Corbusier, Plan Obus

A pair of key notions are to be made, the support concept and the carcass concept. The
support concept, which originated with Le Corbusier's plan libre and Plan Obus, finally
gained full expression in the supports of Habraken and SAR (1965). This concept is
primarily linked with access and the main loadbearing structure and can be characterised,
using Heynen's terms (Heynen 1999), as open and transparent. The carcass concept,
which implies the loadbearing structure if nothing else, is less specifically defined with
regard to the remaining layers and can possibly be best typified as a container or cocoon
that is able to accommodate change. In that sense this concept can be said to represent
seclusion and enclosure.

The frame as a design tool

The frame theory and the accompanying concepts provide a sound tool with which to
analyse the nature and functioning of projects dealing with flexibility or changeability.
The question now is whether the frame concept can yield more than just a tool for
analysing existing projects. Armed with the instruments of analysis I developed, I have
analysed a series of cases, but the question is, can the frame concept be deployed as a
design tool to design houses able to withstand the time factor.

To explore this aspect, I made a overview of every imaginable combination of layers.

This catalogue of frames is then divided among four distinct series of combinations (see
figure 6). These series themselves proceed from a quartet of basic ingredients for
changeable dwellings: basement, carcass, structural wall and facade. It is logical to
assume the loadbearing structure to be part of the permanent component of the dwelling.
7
Two of these four ingredients - structural wall and carcass - accordingly ensue from the
loadbearing structure. The other two - basement and facade - are in that respect deviant
and relate to exceptional circumstances; circumstances which when seen against the
background of recent developments in self-build, can certainly be deemed relevant.
Figure 6. Overview of four series of combinations. Left to right Carcass series, Loadbearing wall series,
Serving core series and elevation series

The series of combinations and their described qualities may be regarded as the basic
material with which to design houses according to the frame concept. Yet many other
deliberations have also to be made during the design process. Taking the dwelling form
and the building typology on the one hand and the desired kind of changeability on the
other, the designer can develop his/her vision and opt for a specific kind of frame. The
frame could exist just out of the loadbearing structure it could be an combination of the
skin loadbearing structure and may be the access system, but also less obvious
combinations are possible.

Figure 7. Collage of the house-core house.

8
For instance it may be possible to create a house based on a frame that consists of servant
elements. Imagine a situation in which a client bought a piece of land, willing to build
him self a house. Imagine again, there will be a firm producing prefabricated elements in
the shape of a chimney containing all the necessities of a house like central heating,
bathroom and a fire place. The client orders such a core of a leading firm. This house-
core is the crystallisation point of the development of a new house (see figure 7). An
important aspect here is that of articulating the frame. While desirable for the frame's
functioning, how does this articulation fit into the concept envisaged by the architect? If
in this case the house-core is well articulated, it claims to be treated carefully.

Here I will put the question of whether the frame concept gives cause for research into
new building materials and constructions and, conversely, whether new materials and
constructions might lead to new potentials for the frame. The above-mentioned
deliberations together with the basic combinations and the combination series, constitute
the tools for designing houses that proceed from the frame concept. It was the intention of
this study to develop the frame concept and the body of concepts attendant on it, and this
I have done in my book Kader en Generieke ruimte (Leupen 2002) (English translation of
the title is: Frame and generic space). Based on a series of case studies this book finally
seeks to be a stimulating resource for anyone involved in developing and designing
houses that are able to accommodate change. The potentials and limitations of the frame
concept can be further explored as designing proceeds.
----------------
Finely I will end with to statements: “Buildings based on the frame concept provides freedom in future
use.” “The stronger the frame is articulated, the more it will claim to be treated carefully, the better it will
be used.”

References
Banham, R. 1969. The Architecture of the Well-tempered Environment. Chicago: The
University of Chicago Press.
Brand, S. 1994. How Buildings learn. New York: Penguin.
Cache, B. 1995. Earth Moves. Cambridge (Mass.): The MIT Press.
Duffy, F. 1990. “Measuring building performance.” Facilities (May): 17-21.
Elias, N. 1990 (1939). Het civilisatieproces, Sociogenetische en psychogenetische
onderzoekingen. Utrecht: Het Spectrum.
Elsdonk, J. v. and H. Fassbinder 1990. Flexible Fixation, The paradox of Dutch housing
architecture. Assen: Van Gorcum.
Habraken, N. J. 1961. De dragers en de mensen. Amsterdam: Scheltema & Holkema.
Heynen, H. 1999. Architectuur en kritiek van de moderniteit. Nijmegen: SUN.
Laugier, M. A. 1977 (1753). An essay on architecture. (oorspronkelijke titel: Essai sur
l'architecture). Los Angeles: Hennessey & Ingalls.
Leupen, B. 2002. Kader en generieke ruimte. Rotterdam: 010 Publishers.
Loos, A. 1962 (1898). Das Prinzip der Bekleidung. Wien/ München.
Ottenhof, F. 1981 (1936). Goedkope arbeiderswoningen. Amsterdam: Van Gennep.
Priemus, H. 1968. Wonen. Kreativiteit en aanpassing.Delft:Technische Hogeschool Delft.
SAR 1965. Voorstellen van de Stichting Architecten Research. Eindhoven: SAR.
Semper, G. 1851. Die vier Elementen der Baukunst. Braunschweig: Vieweg.
Spaeth, D. 1985. Mies van der Rohe. New York: Rizzoli.