Vol. 6, No2 (January 2002)

Eugene, Oregon
Architectural Facts in Search of a Language

I would like to develop some ideas which I think may be important to Neis´ idea of form language. This talk is rather speculative. My hope is to begin a conversation which will help to bring together a few familiar ways of looking at architectural language which are not normally considered within the same framework. Obviously the attempt at seeing architecture as a linguistic system is not new. If I am introducing anything new, it is simply the idea that different points of view each have their own validity, and there is something to be gained by exposing one point of view to the insights of the other. In this way, unhelpful ideological barriers may be breached.

Let me begin with some general observations.

First, my own concerns are with the languages that are shared within a particular culture, and how those languages change from culture to culture and from typical situation to typical situation within cultures. In this regard, I am interested less in the personal rule systems that individual architects follow – if such systems can even be called languages – than I am interested in the formal rule systems that are commonly understood, and which define the architecture of particular cultures. These rule systems may be shared by architects. But my own work has been in the realm of vernacular architecture, and most of the examples in this talk are in vernacular architecture. I define vernacular architecture as including a much broader range of building than only the so-called "primitive" buildings made by non-professionals of relatively unprocessed materials taken from the site itself. My definition has to do simply with what is common, whether or not architects are doing the design. So according to my definition, vernacular architecture does include what we normally understand to be the vernacular, like mud villages in India. It also includes common buildings like office towers in Manhattan, terraced houses in London, houses in Bangkok slums, and prefabricated houses in some parts of the United States. In all these cases, the rule systems for a particular kind of building are shared widely enough so that that kind of building is accepted as the most common and likely way to build.

Second, the reason we need to be explicit about form languages at all is that implicit, shared understandings have almost completely broken down. To be sure, explicit statements of architectural principles go back at least as far as Vitruvius. But the contemporary building world is characterized by an unprecedented diversity of buildings. This diversity is not necessarily unhealthy, but it is accompanied by an almost equally wide diversity of theoretical frameworks. This makes conversation, much less common agreement, extremely difficult. This is perhaps more of a problem in the United States than it is here, but even within the architectural community and the community of architectural education – not to mention the community that includes clients, contractors or bankers – words do not necessarily have the same meaning to different people. I would claim that we need a form of discourse that is understood by more people, and that would lead to more agreement in the judgment of buildings according to the terminology used.

Another way of putting this is that if we are talking about an appropriate form language for our time, or place, or even for particular places, we are talking about shared understandings, and some kind of cultural transmission. I want to make the assumption that such understandings are important, even in the context of an architecture that might include individual expression. The question is that if such shared understanding exist, what is their nature and how can they be most useful.

And third, with this subject we are inevitably talking about the question of representation and of abstraction of things which cannot always be seen. Of course, a building is ultimately a real artifact in the world, tens of meters high and made of stone and brick and mortar. It has a tangible reality, and although the interpretation of its reality might differ from person to person, it exists. Any means of representing it – a plan, a model, a computer fly-through – is only a particular and selective projection of the actual thing. We can be more or less complete in representing the finished thing. But in the final analysis the success of a building is measured not only by its reality, by what it is – which can be represented fairly completely with plans, sections and elevations at different scales – but by the experience of its reality, which I would claim often cannot. Of course, the ability to predict that experience is one of the main jobs of the architect. But even if the architect is ultimately successful in that prediction, what the architect does not necessarily have is a set of tools for sharing all aspects of that experience.

What I want to talk about here is only an incomplete solution to this problem, and perhaps will serve more to illuminate the problem itself than to actually solve it.

Four Approaches to Architectural Analysis

What I want to talk about straddles the boundary between the language of architecture, and the kind of language used to represent architecture. I want to use some examples of vernacular courtyard buildings in different cultures to describe the possible relationships and convergences among four different approaches to architectural analysis. These approaches have stood apart from each other for various ideological and intellectual reasons. Yet, I would claim that in their individual attempts to analyze configurations, they have demonstrated certain similarities of intention, along with the means to together deal with more aspects of architectural experience than any one of them could do alone.

The four approaches are the following.

First, the compositional/typological approach as put forward by the Kriers, for example, in the 1970s and 80s, as the continuation or revival of two centuries of architectural thought, and which still forms the basis for much criticism and discourse. Variations of this approach are used also by various people studying urban morphology in Italy and France.

Second, the approach of space syntax as developed by Bill Hillier at the Bartlett School in London, which has found favor in various urban design and large institutional projects, mostly in the United Kingdom.

Third and fourth, two approaches of Christopher Alexander and his colleagues in California. These include the pattern language, developed during the 1970s, and the latest development in his work which is much more based in an idea of unified form than the earlier pattern language. This new work will be published as part of a four volume work, beginning in a few months.

Each of these four approaches has pluses and minuses, if we are looking for an accurate and comprehensive method of description and analysis, that includes form, the qualitative aspects of rooms, and the experience of buildings. What I would like to ask, is what is to be gained by combining these approaches, so that the positive aspects of one will help to compensate for the negative aspects of the other. Such a synthesis might help point the way toward a form language that has the capability of bringing together various disparate modes of architectural thought.

I would argue that the lack of such a language, that can be shared among architects and between architects and lay people, is one of the major factors that has contributed to the split among various sectors of the contemporary building culture. That split is perhaps more severe in the United States than it is in Europe, but there is no doubt in my mind that architecture suffers from serious problems of communication.

This paper is a speculation, based on observations about these approaches I have made over the last several years. As examples, I am using different variations of vernacular courtyard houses, in different cultures.

Let me begin by trying to describe each of the four approaches, using the example of a courtyard house, built several hundred years ago in Tunis. It is a "classic" example of a house in a North African Islamic city. It consists of a courtyard, open to the sky, and with rooms arranged around the courtyard. The house is entered through a door from the street which leads onto a passage configured so that there is not a direct view of the courtyard from the street. This is to protect the privacy of family life, and particularly of women. There are one or two rooms that do not enter directly from the courtyard, but most rooms do. In addition, the door is the only opening on the street side; all light at the ground floor is obtained from the courtyard.

The Typological Approach

First, the typological approach.

From one point of view, the typological approach is the simplest and most straightforward way to describe this house. In fact, I have already told you most of the terms used: courtyard, arcade, "skifa" which is the passage, entrance, rooms. In describing this house, the typological approach would likely employ a series of simple diagrams that relate this specific example to an ideal archetype. These diagrams show not only the existence of various parts that can be named, but also their relationship to each other. Graphic representation is essential to this approach, as it is to the others.

One reason this approach is so useful has to do just with this simplicity, with the ability for abstraction and the reduction of complex configurations down to simple diagrams that may have abstract meaning. It is these kinds of diagrams that form the basis for the "parti" which we ask students to invent or discover in the course of design. And the parti may connect to certain aspects of meaning. The courtyard, for example, is symbolic of the house as a whole – in fact the words for courtyard and for house are the same in Arabic. And one of the unifying features of North African Islamic architecture is of course that the courtyard building is ubiquitous in the city, serving many different functional types, and this helps to connect the house with the mosque, within a higher understanding that sees the courtyard as an image of paradise.

This approach has a direct and obvious connection to the most common ways of representing a building – namely, the orthographic projections. But as I hope will become clear, this is also its weakness. The orthographic projections only hint at such things as the typical paths of movement through a building, or the differences between two rooms in their feelings of focus and spatial intensity, or some of the connections that may exist between architectural form and human use.

Hillier´s Approach

Let me next describe Hillier’s approach, which emphasizes a different kind of abstraction, that emphasizes a property which Hillier calls "depth." The approach requires that plans be transformed into a kind of diagram that treats all convex spaces as points, or nodes, and doors or door openings between these spaces as lines. What this diagram does is not only allow an immediate picture of relationships among spaces, but allows an understanding of the depth, or number of spatial steps, from one space to another. It is very easy to see, for example, something which is not as immediately apparent from the plan or the typological diagrams – that the rooms that are deepest from the entrance are the ones marked green. And any room may be the root, to determine the depth of all other rooms from it.

The importance of the approach is shown in Hillier’s own analysis of these four hypothetical plans. From one formal point of view the plans are extremely similar. But they are experientially very different. Plan (a) is highly centralized; plan (b) has a kind of enfilade of rooms around the edge, plans (c) and (d) are different variations. The adjacency graphs show widely different topologies among the different examples. One interesting thing here is that the space which appears to be most central, has a very different position relative to the outside in each of the four examples.

Now you may say that this does not really have to do with architecture, and just has to do with where doors are placed. But what I am talking about is the basic topology of configurations, which is connected to the social origins of building type. The power of this approach is perhaps best seen in more difficult cases. For example Hillier has analyzed much more complex vernacular situations to show how power relationships are embodied in the topology of configurations. In this African compound, for instance, the chief’s hut is located in the place that is topologically deepest into the configuration – even though it is physically located near the outer wall.

Perhaps even more interestingly, Thomas Markus has used this approach to study how institutional buildings changed during the eighteenth and nineteenth centuries, to demonstrate how changes in social ideology were reflected in corresponding changes in building plans.

One example shows a school built in London around 1874. This school is an educational hybrid, that combines various features of earlier schools, in which groups of pupils were together in the same large room, with schools that came later, with each group of pupils having its own classroom. Boys and girls had their own entrances, right and left respectively. The plan is quite symmetrical. But the adjacency diagram is remarkable in two ways. One is the very clear gender division, between boys and girls – notice in the plan how the boys’ staircase stops at the first floor – this is node 13 in the adjacency graph. The other is how the groups considered the most vulnerable are deepest into the plan. These are the babies – nodes 5 and 10 – which are very deep into the plan, even though on the ground floor; and the girls, which are very deep into the plan, on the second floor.

Another example is of a workhouse in London built in 1725. The designers of this building intended it to be a model for similar buildings to be erected all over Britain. This is only the ground floor of the building which include kitchen, dining rooms, schoolrooms, work yards and privies. The plan is symmetrical around two axes of symmetry. In his analysis, Markus points out that the rooms which are deepest into the plan and at the end of tree-like structures, are the workrooms and privies – these are the spaces and functions which most needed to be controlled. And the central spaces, which are the spaces from which centralized control was apparently exercised, lie on rings making them accessible to many other spaces of the building.

In contemporary practice, Hillier’s technique is used for analysis of large buildings during the design phase, and it is also used at an urban scale to test different urban design schemes against desired pedestrian activity. One very recent, high-profile use of the technique was in the urban design studies at the time of building the Tate Modern in London, and the pedestrian bridge across the Thames that leads to St. Paul’s.

A major advantage of this technique has to do with the ability to be mathematically precise about spatial relationships, and such precision is necessary when analyzing large configurations . This lends itself to computer analysis of certain aspects of complex projects during the design phase. But this mathematical precision is balanced by a loss of qualitative understanding – a large room and a small room, a light room and a dark room, are all the same in this technique, all other things being equal. But because of its ability to analyze extremely complex configurations, the technique has had success in predicting human movement in cities and such buildings as art museums and hospitals, and in helping to develop an understanding of power relationships in buildings.

I would also point out, before going on, that for the simple courtyard house, it is very easy to put the two diagrams side by side – the one representing the compositional approach and the one representing the space syntax approach – and clearly see their relationship. But as we have already seen, this is not true for complex buildings.

Alexander: Patterns

Let me go on now to Alexander. I’d like to begin with his older formulation, because that is more familiar to many people. The pattern language approach grew out of observations that are essentially functional ones. As such, it was originally a particularly modernist point of view. These functional observations are concerned with recurring human or environmental issues within the environment, and what physical configuration might resolve those issues so that conflict is minimized.

An example of such a pattern is ENTRANCE TRANSITION, which essentially says that the difference between the psychological state one is in on the street and the psychological state one is in inside a house has a corresponding spatial form, in the ENTRANCE TRANSITION, which provides for a change of physical characteristics along a path between street and house. Now there are undoubtedly some healthy situations in which this pattern does not exist or is very weak, and where it does exist it clearly takes on a different form from culture to culture. But in very many situations it does exist, and is even useful because it brings to light what many see as an essential characteristic of houses.

Another example of a pattern is INTIMACY GRADIENT, which says simply that as one goes deeper into a house, one finds rooms that provide more and more privacy and intimacy, and less publicness. Again this has different manifestations from culture to culture; in many Latin American situations the formal living room, where guests are received, is located quite close to the front door; in other cultures such a room is deeper into the house. What is much more rare are houses where such an intimacy gradient does not exist at all, where all rooms are at the same depth from the outside. And for building analysis, the point is not necessarily that this pattern exists or exists in a particular form, but that it is necessary to take this issue into account.

There is a strong relationship, again, between Alexander’s formulation of patterns and those of typology and space syntax. In typological terms the ENTRANCE TRANSITION may sometimes be given the name of a thing: the skifa in the middle eastern house, the fauces in the Roman house, the path-plus-front porch in the small town American house, the outside step-plus-vestibule in the London house. And in Hillier’s point of view, the entrance transition will be represented by one or more nodes connected by lines, the assembly of which gives all of the rooms of the house greater depth relative to the street.

In the case of the Tunisian house, the entrance transition is not simply the skifa; it is the experiential change between street and courtyard that is both facilitated and mitigated by the skifa. Another way of saying this is that if we were to take the skifa alone and build it in the middle of the desert, it would not be an entrance transition, it would just be a set of walls and roof.

Now, looking at diagrams of these three approaches side by side, I want to claim that although these are different representations of the same thing, the emphasis of each one is different enough from the others so that more insight is gained from looking at the three of them than only at one. I will take this up more later.

Alexander: Centers

Alexander’s current work shifts the focus from patterns to what he calls "centers." The emphasis is subtly changed from the relationship between things being at the center of our attention, to the idea of a physical entity itself being the focus, but gaining its intensity and meaning from its relationship with other things. In this respect there start to be some connections to the typological approach. In the courtyard house, the courtyard is the principal center. It is the most important space, and it is the space that will have the most investment put into it in terms of architectural elaboration and ornament. But it does not exist only by itself, but it gains its character and particularly its intensity from the other centers which are around it. Each of those centers, of course, does the same thing, so what we have in a successful building or city is a strongly interacting field of centers, in which the success of any one depends on the success of all the others.

I would like to illustrate this concept further by looking at this carpet, and pointing out that the unity of the composition which exists here comes about because the individual centers have a strong mutual dependence on each other. This is more than simply a figure-ground phenomenon, but has to do also with the strong identity of figures and the fact that such identity exists at different scales.

The idea of centers can help to illustrate typological variations as well. The simple courtyard house is very closely related to others built at the same time. These 27 houses are in Tunis, and all built around the same century. They are arranged roughly in order of size, and probably economic status of their owners. They all share the idea of a central courtyard, rooms around the courtyard, and indirect entrance or entrances from the street. What is happening, as we go from large to small, may be characterized as an intensification of centers. Axes of symmetry that are less important in smaller houses become more important in larger houses, and that importance is reinforced by additional centers, in the form of arcades and rooms which support them. And those rooms, in even bigger houses, are reinforced by rooms which support them and give them even greater intensity.

So in this analytical approach, like in the typological approach, there are entities that can be named. But unlike the typological approach, in which the description of a building is to some extent a description of entities and their relationships, this formulation emphasizes that the life of entities comes, at least to some extent, from the entities around them, and all the entities that are a part of the same composition.

Comparisons among Different Buildings

So far I have used the example of the simple courtyard house to hint at different methods of analysis and representation.

What I would like to do now is introduce three buildings in addition to the Tunis house, all of which may be loosely categorized within the courtyard type, and all probably related to the ancient Roman house, that will provide the material for analysis in the main part of this paper. These examples present different issues of organization and space, and comparison among various groups of them will help me to explain the positive and negative aspects of each method of analysis.

First is the simple courtyard house itself.

Second is a related building, in Cairo. It is the Bayt Suhaymi, a well-known merchant’s house from the 18th century. It too has a main courtyard, and an indirect entrance from the street. It also has an open porch called the taktaboosh, between the courtyard and a rear garden, and a three story, elaborately finished receiving hall off the courtyard, called the qa’a. The indirect entry in both the Cairo house and the Tunisian house is a typological transformation of the fauces in the Roman house. The taktaboosh is a transformation of the tablinum in Roman house.

Third is a courtyard house several hundred miles south of the Cairo merchant’s house, a building that was documented before the construction of the Aswan dam and probably does not exist any more. This is a Nubian house, belonging to a people who were influenced both by the Islamic cultures to the north and the African cultures to the south. In this house the courtyard takes up a larger percentage of the house area than the urban houses we have already looked at. These houses are usually oriented the same way relative to the Nile, they have shaded places for water on the south side of the courtyard, and a special bridal hall for newly married couples. They also had a tradition of being elaborately painted on the outside. In these buildings it is often possible to see through to the courtyard from the outside, and it may be that the much lower density of the Nubian village obviated the need for the courtyard to be hidden. But even here, we may compare the covered loggia with the taktaboosh of the Cairo house.

And fourth is a courtyard house in Mexico, which may be related to the other houses through its own typological ancestors in Spain. Its courtyard is entered through a passage known as the zaguan, similar to the fauces, and has all of its rooms – more public rooms including the estancia which is a living room and comedor which is a bedroom, private bedrooms as well as service rooms – around the courtyard.

I would now like to make various comparisons among these houses which point up the positive and negative features of the various methods of analysis.

Privacy and Function

Let me begin with a comparison of the Mexican house and the Nubian house, from the point of view of issues of privacy.

Although the houses are both courtyard buildings, and share the typological idea that the courtyard is both for circulation and other uses, and also share an entrance room in between the outside and the courtyard, there are many dissimilarities in form. The Mexican courtyard is almost square with the suggestion of an axis of symmetry entering the courtyard from the zaguan. The Nubian courtyard is much more rectangular with no overall symmetry. The Mexican house is one room deep on four sides of the courtyard; the Nubian house is two rooms deep on two sides of the courtyard, one room deep on a third side, and has only a wall on the fourth side. Although they are both clearly members of the courtyard typology, they are formally very different.

But a comparison of the adjacency graphs shows some interesting similarities. First of all the courtyards are each approximately the same number of steps in from the outside – either 3 or 4. Second, the bridal hall in the Nubian house and the parent’s bedroom in the Mexican are the same number of steps in from the outside  – 5 – and are in each case at the deepest level of rooms for human habitation in the plan. And finally, in each case, at least some of the stables are at the deepest or next-to-deepest levels in the plan – no doubt to give protection to valuable animals.

There are several interesting things about the Mexican house. One is that not all the bedrooms are at the same depth inside the plan. In fact, the bedroom in the corner may be the bedroom of the parents – one can see a similar relationship in other Mexican houses. Another is that the kitchen and service rooms are quite deep into the house and this accords even with what we know about more modern houses in Latin America.

Now I do not want to claim that there is a cultural connection that is causing the correspondence in these privacy relationships between the Mexican house and the Nubian house. But what is interesting here is that we have found a method of analysis that can uncover facts that are not apparent with a casual reading of the plan. Of course, a careful reading of the plan alone will yield the insights we get from the adjacency graph. But that is just what the adjacency graph is – a way of reading the plan that abstracts certain of its characteristics. As we have already seen, in the case of much more complex plans, this turns out to be critical.

What has not yet entered very much into this discussion of privacy relationships is use and the function of rooms. Neither the plan analysis nor the adjacency graphs tells us very much about use.

In the case of the Mexican house, the public rooms – the estancia and the comedor, along with the shops – are on the right as you enter, and the bedrooms are on the left. The estancia is also the biggest room and has two pairs of doors, but that doesn’t matter in a space syntax analysis. We have a hint at the importance of this side of the courtyard when we look at the paving pattern, and the size of the estancia – but only a hint.

And it is with these issues, only hinted at with the plan analysis and hardly dealt with by the adjacency graph, that I would suggest that a graphic representation of Alexander’s patterns might be helpful.

The two houses we are considering here, both incorporate the idea of the intimacy gradient, but in different ways, which show that the patterns have different cultural manifestations. As with the typological approach, the way to see these differences may be graphical. In the Mexican house, the existence of the continuous paving from the zaguan to the estancia and comedor suggest that path for the visitor, and make that side of the house seem more public, thus subtly indicating an intimacy gradient which puts the bedrooms deeper into the house, even though they are quite close to the front of the house.

In the Nubian house, there is actually something quite similar going on, where the visitor entering the main courtyard is drawn toward the loggia and away from the wall enclosing the small courtyard. In addition, the door leading to the private courtyard is not even visible from the main entrance to the house, reinforcing the increased privacy of that realm of the house. The private courtyard and the large loggia are at exactly the same depth in the plan, but the way the building is actually made gives them quite a different status in the intimacy gradient.

So the various analyses are saying slightly different things. They are not inconsistent with each other, but each one is emphasizing something that the other is not.

If we look at what Hillier calls "depth", and we look at what Alexander calls the "INTIMACY GRADIENT," and we look at the subtleties in the plans, we see slightly different aspects of just how the architectural space of the buildings are arranged relative to their main entrances.


The Qualitative Character of Rooms

One of the most striking differences among the several approaches is that concerning the qualitative character of rooms and what that does to their role in the overall composition. Let us for example look at the Cairo merchant’s house, and look at it in three different ways: from the point of view of a plan analysis, from the point of view of space syntax, and from the point of view of a field of centers in Alexander’s formulation.

The room called the qa’a is on the side of the courtyard, not on its main axis, but perhaps in a position that corresponds with the alae or sometimes the triclinium in the Roman house. Indeed, there is another interesting correspondence with the Roman house, and that is that the taktaboosh, which is the porch in between the main courtyard and the garden, is very similar in position and function to the tablinum in the Roman house. This building was built about 1300 or 1400 years after the fall of Rome, so comparisons are dangerous. But if we look at this plan, it is as if the symmetrical plan of the Roman house was relaxed, the direct axis in from the street disappeared, the shops on the street went away, and that was the nature of the transformation. And of course that accords with what we know of the social forces which helped to shape the Islamic city.

But getting back to the qa’a. The adjacency graph of the house does not say much. It is one room among many and its adjacencies are similar to those involving less important rooms. If we look at a plan analysis of the house, where we look at the relative size and symmetries of rooms, the qa’a begins to assume more importance, since it is one of the more symmetrical large rooms of the house and adjacent to the courtyard.

This importance is reinforced when the plan is extended to a section or perspective, since the qa’a is a three-story high space.

But there is something that even the plan or the axonometric diagram does not emphasize about this space, and that is the fact that it has received the most investment in terms of materials, ornament, local symmetries, and a major lantern in its roof which brings in light. It has a very high level of architectural intensity, in my experience of the building rivalling even that of the courtyard itself.

This may be represented by a different sort of diagram, which is a simple diagram that illustrates the distribution of centers and their intensity. With present day computer techniques I’m not sure if this diagram could be analyzed as easily as the adjacency graph. But it does immediately call your attention to the qa’a, and immediately indicates the intensity of ornament and architectural elaboration that may be present there.

Compare this with the estancia in the Mexican house. The estancia is similar in that its importance in the life of the house is only hinted at by the plan or by the fact that it is the only room with two doors toward the courtyard, and the only room which has niches built into the walls. It may have a higher ceiling and more architectural elaboration, and it will almost certainly be felt as the most important center of the house, of equivalent importance to the courtyard itself. And this status would show on a mapping of the centers of the house.

Before I go on to the conclusion I would like to show these drawings and corresponding analytical diagrams from Rob Krier’s book on architectural composition. The diagrams are intended to illustrate principal spatial characteristics of these buildings, by giving emphasis to their major spaces and spatial sequences. At least that is how they are described in the text that accompanies the illustrations. The text does not make any mention of any differentiations within the spaces themselves, no mention that in particular situations, for example, a wall or a column may have an increased intensity or sense of focus. But in fact the way the diagrams are drawn is suggestive of something which is not mentioned in the text, and that is that within particular spaces, there may be an emphasis on material or ornament or the surfaces themselves. In other words, these diagrams are suggestive of the diagram I drew for the qa’a, illustrating how the formulation of Alexander’s centers may help to interpret, in an abstract way, aspects of buildings that may not be immediately extracted from the orthographic drawings.

Conclusion and Applicability to Contemporary Work

Within each of the methods of analysis it is possible to compare one building to another, within particular types.

Certainly the plan analysis itself does this, and individual examples are often described as variations on an archetype. This is commonly done in studies of architecture and vernacular architecture.

Bill Hillier and Julienne Hanson have shown, very rigorously, that a space syntax analysis of buildings within a particular culture yields correlations among their adjacency graphs. So there is, in a sense, a typical adjacency graph for certain kinds of buildings within particular cultures.

In her work on the Roman house, Carol Watts has shown, equally rigorously, that there is a recurrent set of patterns that is characteristic not only of the domus but of the insula as well, showing the persistence of patterns within a particular culture over different conditions of density and formal typology.

Such a study has not yet been done for Alexander’s formulation of centers and fields of centers. But the idea that there might be culturally-specific features in distributions of centers was hinted at in a dissertation done by Artemis Anninou. In teaching about vernacular architecture, I sometimes refer to particular building types in particular cultures as culturally specific and culturally shared fields of centers.

I am describing all this to make the point that all each of these ways of looking at architecture has its own rigor that stands up well to detailed analysis. In this paper, I have rather less rigorously shown that that on the one hand, these approaches are consistent with each other, and on the other, they bring to light different but equally important aspects of buildings.

So just to summarize:

The well-known typological and plan diagram approach helps us to understand the features of buildings that are related to formal archetypes, and which are sometimes connected to the global meaning and symbolism of buildings.

The space syntax approach helps us to accurately understand the detailed topology of configurations, which is often connected to how buildings and cities reinforce their meaning through power relationships, social distribution and human movement.

The pattern language approach helps us to see the relationship between formal relationships in buildings and human use and function.

And the approach of centers helps us to understand the actual spatial presence and intensity of rooms, groups of rooms and other entities in the building.

I have tried to argue, first, that each of these approaches has its own strengths and weaknesses; second that they are nevertheless compatible with each other; and third, our understanding of buildings is increased greatly if we use all of them together. The title of this paper, "Architectural facts in search of a language," refers to the fact that there is not a common language of representation that includes them all.

Such a common language might be helpful in contemporary work, and particularly to the development of complex configurations such as urban design schemes, housing, and large civic and institutional buildings. Such projects need to have architectural beauty and coherence, but they also need to respond very specifically to issues of human movement, function and feeling which are often extremely difficult to understand when looking at designs done only with typical representational techniques.

Let me just show, without analyzing them, two buildings which were published about five years ago. This is an old persons’ home in Finland; the other, which I will show in a minute, is a center for handicapped children in France. Both of these buildings are highly complex, and as buildings for communities of people that are vulnerable in different ways, they need to balance personal architectural expression on the part of the architect with a high level of understanding of how the building helps or hinders human encounters of different kinds, helps or hinders the reinforcement of self-identity and self-worth that vulnerable people need to feel, helps or hinders the healing that a well-designed room might contribute to.

The old person’s home is described in an article about it as fulfilling a need to have "a clear, economical and reassuring overall pattern, and at the same time a gentle gradation of privacy from the public route, through the semi-private space, to the privacy of the individual space." And indeed, we might believe from the plan that this intention has been met.

The children’s center is described as having "a key area [which is] the main hall or common room, which is open to everyone and where local people, notably the elderly and adolescents, are encouraged to drop in. Conceived as a place of stimulating contact and exchange between handicapped children, parents, friends, visiting school children and the local community, it is designed on an open plan with an island fireplace, and contains such attractions as a multi-screen TV wall, a small lending library, an aquarium and a musical chequer-board floor." And again, we see in the plan and photograph a space that might have the life that is described.

These are both expensive buildings with ambitious social agendas. There is no doubt that with these buildings, there was a close relationship between architects and clients, and that presumably the client group felt that they were in a position to criticize the design based on their best knowledge of their own specialties of social service or medicine. But what we have seen in this talk is that not everything that needs to be known about buildings can be learned from a simple reading of the ordinary design drawings – and so that even an enlightened architect and client, working together with the best of intentions, are not necessarily in the best position to predict success.

And although it seems as if the human and social needs expressed in these buildings may only be important for vulnerable groups or groups that need to be controlled in different ways, like children or victims of Alzheimer’s disease, I would argue that in fact vulnerable groups are like the canary in the coal mine for all of us. The failure of much public housing and urban design often comes about because of a failure to make accurate predictions about such matters as the relationships between public and private space, or pedestrian densities, or the use of open space. Over the last 10 or 15 years, these failures have resulted in a renewed attention to the social aspects of design of complex configurations like neighborhoods and housing schemes.

I would suggest that for such projects, all four of these techniques be used for analysis during the design phase. They emphasize different but essential issues in buildings and building groups, and they offer the promise of representational systems that can allow for good communication, and for success and failure to be understood and shared. The use of contemporary digital techniques means that this should not be as outlandish as it sounds. Systematic analytical methods, along with traditional architectural judgment, would reinforce each other.

The language I have been talking about is a language of design and configuration that is not always obvious in usual techniques of architectural representation. Some of the ideas are embedded in the culture and difficult to find at first sight in the architectural drawings. Because they are difficult to find, they may be neglected in design, and sometimes are. My point is simply that if we were comfortable with simple ways of representation that could go along with aspects of culturally-influenced language that might otherwise be hidden, buildings might gain in their social and cultural appropriateness.