By karenmmartin on situated
Bill Hillier and Adrian Leaman
On a broad level this paper, published in 1973, outlines a theory of the relationship between physical and socially-constructed space. It describes a cyclical process of continual adaptation and reconstruction of existing prestructures and the physical artefacts founded on these prestructures that (I believe) might be used to understand both the design process and the relationship between designer, responsive architecture and audience.
Hillier and Leaman suggest that the environment in which the designer is situated has the greatest affect on the outcome of the design process. And that to understand the design process or ‘method’ as they call it, it is more important to study this environment than to study what the designer actually does.
They break down the factors influencing designers’ decisions into two categories that they call g-models and p-models. These are generally analogous to genotypes and phenotypes but for the fact that they are associated with the quicker moving cultural area than the relatively stable biological field. G-models refer to the underlying prestructures that influence designers’ decisions – often unconsciously – while p-models refer to the site-specific decisions made by the designer concerning the particular constraints and affordances of the specific problem.
The following example is given: “An army marches all day. At nightfall a halt is called and unpacking begins. Within a short time a structured environment appears. Tents of various kinds and sizes are placed in certain definite relations; kitchens, sentry posts, flags, fences and other paraphanalia are erected. A complete environment is, as it were, ‘unfolded’.
The army experiences this as a simple, repetitive procedure. But looked at scientifically in terms of the structures which must exist for such a simple ‘unfolding’ to be possible, it is both complex and illuminating. First, the observable environment which is unfolded is based on a set of instructions in the army manual or in standing orders. Secondly, these abstract instructions and relations are ‘embedded’ in the items the army carries about with it to manufacture this environment, which may be called its ‘instrumental set’. Third, and most important, it will be noted that any series of camps unfolded by the army will exhibit both similarities and differences. In this simplified and artificially deterministic example, the source of similarity is the set of instructions as embedded in the instrumental set, and the source of differences is the local constraints and contingency, which will include personal and environmental factors, strategic considerations and so on.”
In this example the g-model is the set of rules carried by the army and the p-model is the individual instances of the camp. Hillier and Leaman note that an important property of g-models is that they become unconscious or autonomic, that is, essentially invisible to the person acting on the basis of them. Critically, these models also act ‘not only as a basis for creative action, but also as a basis for understanding and interpretation’ which leads to the situation where the g-model becomes ‘both the structure by which the designer analyses and identifies as well as solves his problem’.
Leading on from this example and analysis it is suggested that there are two distinct areas of study when examining the process of design. These are “to characterize the autonomic prestructures by which the designer interprets his problem and which also act as a ‘solution field’; and to characterize the operations which may be performed with and upon such structures in a more or less complex problem environment to produce unique and effective solutions.”
Moving on to examine the structure of the g-model or ‘solution field’ Hillier and Leaman begin by making a distinction between the formal structure which relates manipulable elements – words, walls etc – to other, similar elements, and the less formal relationships which connect these elements to dissimilar elements – a relationship which can establish the social significance of the elements.
Structures come about through the relationships between these elements, and systems are created through combining structures. The key to understanding these systems lies in ‘discovering how the internal autonomic structure of the ‘simplest structures’ of the morphology already contains the rules which govern aggregation into higher logical forms.’ Hillier and Leaman suggest that general system theory has been unable to develop more than a simple characterization of complex systems because the theory presents a definition of a system as ‘elements and their relations’ while the authors suggest that ‘there simply are no elements’.
These relationships between the formal and social structures surrounding each ‘simplest structure’ are described as having the form of a commutative square where ‘transformations in one domain carry with them transformations in the other if ‘meaning’ is to be retained’. The architectural example of a wall is given; “a wall, by reason of its purposeful existence distinguishes at least two types of space… Moreover, if the wall is perforated or not, then it also defines a relation between the two types of space… This is the simplest phenomenal form… To this must be added the relation to a similar domain of such structures, and dissimilar domains of usage and semantics to give the ‘simplest structure’ of architectural form.”
The mapping structures connecting these dissimilar domains are common in human cultures. Within design it is suggested, ‘these mapping devices are used as autonomic devices to solve problems’ while in research ‘these mapping devices are studied in order to understand and improve them’. Could this be the basis for the difficulty in reconciling research and design that I am struggling with? That one seeks to solve problems, the other to understand them?
The next section deals with the idea of the ‘manifold’ as a way of describing the connections between the systems for constructing and interpreting structures. Hillier and Leaman suggest that these connections make it impossible to deal with entirely abstract concepts of space and activity (in architecture) as well as it being impossible to separate abstraction and reality (in this case, the social meaning of a space from its physical form). They go on to say that these ‘manifolds’ (a mathematical, topological term) which make up the designers prestructure is also influenced by the personal history and experience of the designer and ‘the influence of the method of representation itself’ i.e. that the way in which the design is realised, the form, medium, technology available to the designer both practically and conceptually will influence the outcome.
An important property of the manifold is that when part of the manifold is in use then the whole of the manifold has been activated. Although there are two levels of structure – the rules or g-model and manipulable set or p-model these are inter-dependent. Hillier and Leaman use language as an example, saying that; “in carrying out a conversation, the whole language structure is in action both in order to interpret what others say, as well as to plan one’s own.”
They then move onto discussing the transformations it is possible to affect on this underlying genotype structure. They equate this to the question of which time structures are relevant to design, saying that there are two levels at which time is critical to an understanding of design. These are ‘evolutionary time’ and ‘local time’. Evolutionary time “includes the development and stabilization of technologies, social processes of environmental control and all constructions of the mapping from social into spatial form.” It is this that has the greatest significance on the design process by affecting the underlying prestructure which designers bring to each particular design problem in ‘local time’. Local time is defined as the time taken for individual designers and users to adapt the g-model into an appropriate p-model for a particular instance.
Four types of local time structure within design are identified:
- The process of variety reduction where an increasing number of possible solutions become excluded as time goes on leading to an increasing specification of the problem and the convergence of ideas on a unique solution.
- The process of conjecture test where potential solutions are simulated and tested against the particular situation (p-model).
- The process of communicating the design process through consultation and production of communications tools and designs. This should not be confused with the underlying design process though it may influence the result.
- The process of mapping between the g-model and the p-model. This may be done using a generalized form of the conjecture test process.
It is this fourth type of time structure which Hillier and Leaman focus on. They suggest that the process is analogous to that of mapping between the abstract understanding of language and verbal production; “The speaker, like the designer, starts from a prestructure in which the most important entity is the abstract structure by which mapping between dissimilar domains may be effected. In all such cases there is a manipulable set. In the case of language, the manipulable set is words and phrases; in the case of the designer, it is the representations of the instrumental set. Without the mapping structure these are virtually useless and only meaningless sounds or arbitrary artefacts could be generated.” This results in a two-stage process; for language, stage one is the mapping between the dissimilar domains of abstract, structured meaning to a general idea of words and structure for speech. The second stage is the mapping of this generalized idea of speech into a precise verbal structure which will be spoken. In design the two stages can be identified as firstly, the mapping between the dissimilar domains of the statement of the problem to a general solution, stage two is the mapping from a general to a specific solution. They conclude that science also uses this conjecture test method and so “it appears that this simple idea can unify speech, design and science as being transforms of the same basic cognitive activity.”
Two different approaches to the conjecture test design process are described through the metaphor of a detective at a crime scene in search of the culprit; “The crime and the villain may be imagined as the end point of a chain, the links of which are the clues in a deductive sequence. The detective may either conjecture a villain and attempt to construct the clues in such a way that they lead to him. Or he may try to build the chain of clues from the crime ‘outwards’, in the hope that they will lead to a unique villain.”
The conclusion of the paper returns to architecture.
“If design is to be understood beyond the mistaken slight phenotype modification for the continuous recreation of a genotype, then observers must become something like ‘archeologists of the present’ attempting to uncover the elementary forms by which our societies are being mapped into new spatial structures… The first question for a theoretical science of the environment is ‘what social behaviours determine the environment?’”
They continue; “In fact, the design model outlined above may be continued through the user who, like the designer, takes a prestructure, elaborates it, discovers it, and cognitively and actively transforms it.”
Hillier and Leaman conclude that complex structures and theories develop through the combination of elementary structures. These elementary elements include both physical and social structures in an on-going cyclical development process. These ‘simplest structures’ already contain the rules “which give the spatial structure characteristics of urban and other higher-order spaces, as mappings of social processes. These will lead, by further elaboration, to different forms of structural stability to which both social and physical factors contribute.”
While the approach outlined here draws many analogies between language, speech and design, coming from a dance background I am interested in how non-verbal forms of communication might also fit (or not) into these analogies.
I don’t have a deep knowledge of this area but it seems to me that the theory outlined in this paper has its roots in semiotics and structuralism. Structuralism is described within the paper as proposing a strategy for studying social systems which reflects that used in modern theoretical physics which, the paper says, accepts the basic duality that ‘on the one hand it investigates formalisms; on the other, it investigates the physical universe itself, making models in the former which satisfy the appearances of the latter.’ I wonder how much this underlying theory and approach influences how design is viewed?
If science and design share a similar process (though very different methods for evaluation perhaps?) is the distinction between science and design an artificial one? In this case, the creative process might be applied to a far wider field than merely ‘design’. And while I recognise both of the ‘crime scene’ processes as design methods, I suspect there may also be a less ‘deductive’ process possible where the goal is not this clearly defined beforehand.
* Keep New:
* Posted on: Mon, Aug 7 2006 1:11 PM
* Updated: Mon, Aug 7 2006 1:24 PM
* Email This
* Clip/Blog This
