ARCHITECTURAL TEMPERATURE AS AN ABSOLUTE MEASURE IN DESIGN, by Nikos A. Salingaros
© Nikos A. Salingaros, 1997. Submitted for publication in the Journal of Architectural Education.
Abstract
The architectural temperature T measures the degree of detail, curvature, and color in a design. Every building can be characterized by its value of T . The greatest historical buildings have a high value of T , comparable to the T of the human form. This paper argues that human beings connect only to objects that also have a high value of T . One seeks to achieve a state of equilibrium in T with one's surroundings. Our results are based on fundamental processes in physics, biology, and the theory of complexity, and are supported in part by findings in experimental psychology. This raises important questions about the impact of buildings, which can now be investigated in a quantitative manner.
INTRODUCTION
Architectural design has always benefited from a scientific analysis. There have been serious efforts in the past to find mathematical truths in terms of the Golden Ratio and the Fibonacci sequence, though the basis for design remains rather mysterious [1]. Architects still have to treat their field more as an art than as a science. Even so, there is purpose in looking for a scientific basis to architecture. Scientists can help in this task, even though there are formidable complexities involved in analyzing architecture mathematically.
This paper adopts a particular approach that helps to make a scientific analysis possible. We view architecture as a perceptual connection between forms and the human mind. Instead of regarding forms as existing in isolation, we focus on the perceptual connection process. In this, we are guided by quantum mechanics, where the observation itself plays a deciding role in determining the physical state. We then use an analogy with thermodynamics to model the main features of the interaction process. The model is further developed into a predictive tool, and its predictions are in large part verifiable.
We propose that a largely subconscious comparison mechanism plays a crucial role in the perception of our surroundings. Architecture depends on the interaction between a design or surface and an observer. This process measures information about the environment. The information is processed in the mind, where a continual comparison is taking place with built-in fixed internal scales. We can take conscious action depending on the result of this comparison, otherwise the effect is subconscious. In either case, this interaction generates psychological and physiological responses that determine how the structure affects us.
We introduce the architectural temperature T to measure the degree of detail, curvature, and color in a design or surface [2]. T can be easily estimated on a scale of 0 to 10 by breaking it up into five independent qualities:
- sharpness of localized detail;
- density of differentiations;
- degree of curvature;
- intensity of color hue;
- contrast among different color hues.
Every building has its approximate value of T . One may rank-order buildings that belong to different styles, and from different epochs, based on their value of T . We do this for twenty-five buildings representative of the various periods of architecture, later in this paper.
Our own response to a building depends on the value of T . A human being connects with a design or surface proportionally to the information it contains. As the information is proportional to the architectural temperature T , it follows that the degree of connection is proportional to T . Therefore, measuring T using a set of rules gives us a rough estimate of how strongly we connect to that design or surface. We can predict our response to a form by means of a few simple measurements, and this is verified in practice.
There exists a reference value, or range of values, for the architectural temperature. The human mind has developed over the last million years or so by processing information from its surroundings. It seems reasonable that the inherited patterns that define an internal reference T should have values corresponding to the natural forms that played a role in our evolution. This assumption leads to a value for the reference T as being somewhere between 7 and 10. Our internal T is therefore the same as our external T (corresponding to the human form itself), which is comparable with all other living beings.
We conclude that human beings connect more to high-T forms and less to low-T forms. This statement refers to psychological and physiological responses based on a neutral model, and has nothing to do with personal preferences. One's opinion of what constitutes a beautiful form is developed through conditioning - learning by example - and is not directly measured by T . Our model deals with basic mechanisms that lie beneath what we are taught to consider beautiful or elegant. This response appears to be inherited and not learned.
Note: Throughout this paper, the symbol T and the word temperature refer to the architectural temperature, and not to the thermodynamic temperature.
THE CONNECTION PROCESS WITH SURFACES
Human beings connect to surfaces and shapes through structural subdivisions and color. Information is continually coming in from our surroundings and connects us to a building [1]. The comparison process of design information between observer and environment may be described in thermodynamic terms. Heat is exchanged when the thermodynamic temperature is unbalanced, through a connection established by conduction, convection, or radiation. Temperature differences set up a heat gradient, and the resulting forces neutralize those differences whenever possible.
We propose an analogous process involving the architectural temperature T . Differences in the architectural temperature between high-T and low-T regions set up forces that are experienced in our mind. What affects us the most is differences between the T of our surroundings and our own internal reference T . We are able to establish equilibrium only with a high-T environment. Being ourselves high-T objects, we connect strongly when the T value of a building balances our high internal T ; otherwise an imbalance remains. An observer might find a low-T surface intellectually "pure", and be excited by its strangeness, but he or she can never establish equilibrium with it.
A RESULT FROM EVOLUTIONARY BIOLOGY
Results from the new science of complexity help us in understanding the theoretical basis of the architectural temperature from a different perspective. One result that is directly relevant here concerns criticality in the creation of self-replicating organic molecules. In the simplest models, a certain minimum of different types of organic molecules in an autocatalytic cluster is needed before the evolution of life can take place [3]. This works as follows.
Computer simulations try to establish a self-sustaining group of molecules, which would in theory give rise to life forms. The simulation proceeds by adding different organic molecules, such as one would find in a primordial organic "soup" on the early planet. We expect that some of these molecules will catalyze chemical reactions involving the others, in what is called an autocatalytic set of molecules. At a threshold number of diverse molecules, a giant web of catalyzed reactions results in a phase transition [3]. Catalytic closure is achieved through diversity.
This result is important by analogy. If we don't have a sufficient number of distinct architectural elements, the architectural temperature of a building is too low to make it connect to the human mind. A coherent ensemble needs to have the requisite number of different forms, as measured by the architectural temperature. The nuclear chain reaction is also illustrative, but not entirely analogous. A subcritical reaction extinguishes, whereas a supracritical reaction leads to an explosion. A critical mass (of the same type of Uranium atom) is necessary for the chain reaction to self-sustain.
PHASE TRANSITIONS AND ORGANIZATION
There is a further direction that the model of this paper takes us. Suppose we start with an amorphous surface at zero architectural temperature. We then continue to add design elements that raise T in steps. As argued above, this will lead to a structure with more information. In most cases, it will be perceived as more interesting, though this depends on other factors as well. Whether it is actually preferred is still determined by prior conditioning in the observer. Continuing to raise T will eventually lead to a design that is too rich or disordered, and will be difficult to grasp conceptually.
What happens next is clear if we look at the thermodynamic analogy. An initially amorphous state that is heated will combine (melt) together. If we now allow it to shed some of its energy, it might well jump into a state of greater order by crystallizing. By undergoing a phase transition, a liquid will give off its latent heat of fusion, and assume a vastly more ordered state. The important feature of crystallization is that it occurs at a fixed temperature. Ordering re-arranges the basic components without lowering the thermodynamic temperature.
An analogous process occurs in architecture. One can raise the architectural temperature of a building or surface, and then impose an ordering through symmetries and connections. By this expedient, we keep adding color, curvature, and detail in a very controlled and precise manner, while T stays at its high value. An excess of architectural information is shed by means of the imposed ordering. The degree of ordering is measured by the architectural harmony H [2]. A hot disordered state has high T and low H , whereas a hot ordered state has both T and H high.
We will not describe this ordering process here, even though it is central to any theory of architecture. The model can be expanded to include the architectural harmony as well, and that is discussed in detail in Ref. [2]. As is evident from the list of T-values of buildings given later, there is a coincidence between buildings of different quality that have the same T-value. In physics, that is a clear indication that the phenomenon being measured depends on another independent quantity. Measuring that other variable breaks the degeneracy, and separates states that originally coincided.
THE ARCHITECTURAL TEMPERATURE OF BUILDINGS
T takes values on a scale of zero to ten. Each of the five components of the architectural temperature listed at the beginning of this paper will be given a rating of zero to two, and all the five scores then added together [2]. Assign a zero if there is no attribute; assign a one if there is some, and a two if the quality is present to a significant extent. This is the easiest possible measure of the architectural temperature T , and is a highly simplified means to handling an extremely complex topic. A moderately high T can be achieved via many different combinations of the five factors.
In the following Table, we have estimated the architectural temperature T of some well-known buildings [2]. The numbers obtained are very approximate, and are meant only to demonstrate proof of principle. Buildings are listed in decreasing order of T , with the earliest buildings listed first within each group having the same T values.
Building (Architect), City
T
Hagia Sophia, Istanbul
10
Alhambra, Granada
10
St. Peter's, Rome
10
Taj Mahal, Delhi
10
Watts Towers (Rodia), Los Angeles
10
Dome of the Rock, Jerusalem
9
Hundertwasser house, Vienna
9
Konarak Temple, Orissa
8
Maison Horta, Brussels
8
Casa Batlló (Gaudí), Barcelona
8
Parthenon, Athens
7
Palatine Chapel, Aachen
7
Phoenix Hall of the Byodo-In, Kyoto
7
Cathedral, Salisbury
7
Baptistry, Pisa
7
Carson, Pirie, Scott store (Sullivan), Chicago
7
Medical Faculty (Kroll), Brussels
7
Pompidou Center (Rogers), Paris
6
Fallingwater (Wright), Bear Run
4
Opera House (Utzon), Sydney
4
TWA Terminal (Saarinen), New York
3
Shanghai Bank (Foster), Hong Kong
3
Notre Dame du Haut (Corbusier), Ronchamp
1
Seagram Building (Mies van der Rohe), New York
1
Salk Institute (L. Kahn), San Diego
1
Every building in history can be assigned an approximate value for T . This allows buildings to be rank-ordered according to the value of their architectural temperature. Our arguments support high-T buildings, yet many people prefer low-T buildings. In any case, this is only one of several architectural qualities that characterize a building. As already pointed out, T does not measure what one personally likes or dislikes. Neither can T be used as the sole criterion for judgement. Though they share the same T-value, the Watts Towers by Simon Rodia certainly do not rank with the Taj Mahal in greatness.
Non-scientists have unrealistic expectations that a single variable should perfectly describe a complex phenomenon, and discard a model when that doesn't occur. In the present case, an architect might well object to a model that ranks the Taj Mahal with the Watts Towers. But that would be to misunderstand the scientific method, which applies a rough model to all possible situations, carefully noting its varying degrees of success. This information serves as a basis for expanding the model to include other variables. A model involving several variables has to be built up in a sequential fashion, especially when all the variables are not known at the beginning.
APPLICATIONS TO ARCHITECTURAL DESIGN
The more things that are happening on the small scale, the more information there is in a design. Details, contrasts, and curvature all connect to the observer in a positive way up to a certain point. One may like either colorless, colorful, plain, or detailed designs; yet their impact is independent of personal preference. Beyond a certain point, however, even those who prefer lots of detail and color will feel that it has become too much. Just as with a thermodynamic system, a set of objects will become disordered with rising temperature. In an architectural context, disorder makes a viewer uncomfortable.
Our perception of the natural world is based on the ability to grasp ordered and not disordered complexity [4]. T measures architectural information: the process of pattern recognition taking place in the mind of the observer is helped by spatial ordering, and is hindered by randomness. The mind can process information from the environment much more readily if that information has some ordering. As living beings have a high degree of architectural temperature, we have also evolved techniques of grouping high-T designs into regular patterns. This ordering makes an enormous psychological and physiological impact on a human being.
Throughout human history until the 20th century, great buildings have resulted from increasing the architectural temperature. This is independent of what the actual buildings look like. Adding small-scale structure that doesn't relate to the ensemble, however, lowers the harmony of a design. Buildings characterized as "too rich" have excessive decoration that makes an overall coherence impossible. It is not the quantity of detail that makes a great building, but rather how successfully each element correlates with the rest of the structure.
Natural materials and surfaces have an intrinsically high T . They are used by traditional as well as contemporary architects. Though this practice is widespread, however, it ignores the dependence of high-T designs on the distance to the observer. The architectural temperature of a surface should be high regardless of the distance, but natural materials are high-T only close-up. The details can be seen effectively within a limited distance. Farther than that, the fine structure that raises T is lost, and the surface becomes low-T [5]. On regions farther away, differentiations have to be much larger so as to appear the same size as small detail would be at arm's length [5].
Human beings cannot connect to a building that has zero architectural temperature. The commercial sector knows this very well, and attracts customers by using techniques that raise T directly. A large portion of the advertising industry is involved in raising T in the built environment through billboards, signs, lights, etc.; competing to draw people to higher-T objects. It took some time for architects to accept this as an architectural phenomenon [6, 7], and we continue to artificially separate the building industry into low-T buildings and high-T signs.
THE ARCHITECTURAL TEMPERATURE OF NATURAL FORMS
The ultimate validation of this model is that it also applies to natural forms. It is possible to compare buildings directly with natural forms, using the scale introduced above. That provides us with reasonable estimates for T . The architectural temperature due to geometric differentiations is directly related to the fractal geometry of nature [8]. For example, weathered rocks have a fractal structure, which gives them a T value from 6 to 8. Trees and plants also have a fractal structure, and usually rate an 8 or 9. For a blossoming plant with flowers, or a tree in the fall with colored leaves, the value of T rises to 10.
Animals and people can be judged on the same scale. Most mammals rate a 7 or 8. Brightly-colored birds and insects can rate up to 10. The architectural temperature of a human being is of course strongly dependent upon clothing. A man in a business suit rates a 7, whereas a woman in a colorful folk costume rates a 10. Living creatures all tend to have a T value anywhere from 7 to 10, and that establishes a natural range. Our model makes the assumption that a reference value for T has evolved as part of the human consciousness.
Man-made objects that are not buildings can be rated according to their value of T as well. Colorful oriental carpets can rate up to 10. Other traditional woven textiles, such as from Guatemala, Indonesia, and Japan, also rate anywhere from 7 to 10. Wooden carvings from India and Southeast Asia rate from 7 to 10. This list can be extended indefinitely to include all objects that man finds pleasing. Notice that all over the world, it is precisely such high-T objects, either created or natural, that connect with a person, and this is the reason they are valued. The predominance of artefacts with high architectural temperature confirms that they satisfy a profound internal need in human beings.
INTERPRETING THE INFLUENCE OF STYLE
The above results should encourage one to design high-T buildings, consistent with (though not copying) the great buildings of the past. That goes against the architectural tradition of the 20th century, which tends to be based more on stylistic concerns. For example, a minimalist style eliminates differentiations altogether, and seeks a T near zero. Anything that is uniform, colorless, without differentiations, and using amorphous materials achieves that end. Modernism arose in opposition to traditional architectural styles - which all have a relatively high degree of architectural temperature, independently of what they look like - and is therefore a low-T style.
Some famous modernist buildings use beautiful natural materials such as polished wood or stone. The small-scale natural details in the surface introduce differentiations on the smallest scale. This practice, however, disguises the minimalist intent of the style. In general, natural surfaces used in modernist buildings are combined with severe straight lines and a lack of substructures so as to assure a low value for T . Although at times both color, and curves, are used, the stylistic hallmarks of modernism are the straight line and the right angle. Where color is used, T is lowered by eliminating details and curves; where curves are used, T is lowered by eliminating details and color.
People have always been fascinated by unnatural images. An object that stands out because it is completely regular, or shiny, or transparent, fascinated early man, who worshipped golden statues, meteorites, and monoliths. We sometimes build temples that have an unnatural shape by choice, such as the pyramids. These express our mastery over nature in a forceful and grandiose manner. Today's culture celebrates strangeness, and is bored by what is natural, and this fascination helps to keep the architectural temperature of contemporary buildings away from equilibrium.
An architectural style can connect to people by endowing intrinsic mathematical qualities to the man-made environment. Alternately, a style could deliberately detach human beings from buildings by emphasizing the unnatural qualities of design. Our model provides a useful tool for understanding this process.
APPROACH TO EQUILIBRIUM VIA WEATHERING
According to our model, there are two methods of establishing equilibrium in the architectural temperature. The first is to do so deliberately - adjusting the T of a design so that there is a balance with our own internal T . If this process is not invoked at the time of construction, then the much slower process of natural weathering starts to take place. On exterior surfaces, natural changes in the materials generate structural features that will raise a low T [5]. This latter process, however, occurs only where a building is exposed to natural forces, and then the nature of the materials determines the time frame necessary to achieve equilibrium.
This observation raises issues on the nature of materials that cannot be analyzed here. An important aim of certain architectural styles is the search for a low-T material that maintains its low-T character against weathering. In this author's opinion, that is an ultimately futile quest, going as it does against natural processes. Low-T surfaces cannot remain low-T . The basic notion of stability in physical systems shows that states are long-lived only if they do not have to be propped up - if their energy is such that all inevitable small changes reinforce the state instead of disturbing it drastically.
EMOTIONAL IMPACT OF THE ARCHITECTURAL TEMPERATURE
Although this paper is not directly concerned with environmental psychology, there is a relation between the connection process discussed here, and the degree of emotional comfort one feels next to or inside a building. Regardless of their style, how well buildings work can be judged solely on the basis of their connection to people. We interpret the emotional response of a building as due in part to mathematical qualities in its design. Different values of T will lead to different degrees of emotional comfort.
Studies tend to confirm that high-T structures are emotionally nourishing [9]. Of course, this depends on many other factors as well, yet the basic effect is clearly there. We can supply a reason for why different values of the architectural temperature create either a positive or a negative emotional response. Our mind is continuously analysing the information content of designs and surfaces, so the process depends on the length of time we spend at or near a particular building. An artificially empty or minimalist form frustrates our perceptual system, which is seeking to extract information from the environment at all times. This effort uses up energy, and this can be perceived as an emotional drain.
Many people feel such things instinctively: they try to raise T in their surroundings by having things they like around them - objects with a rich color and detail, or certain colors, curves, and textures. If they can afford to, they surround themselves with beautiful antiques. People without money will paint their house in bright, vibrant colors, and put up inexpensive but effective decorations (unless they are influenced by a minimalist style). The need to raise T may be interpreted either as an avoidance of the perceptual energy spent on low-T objects; or as the pursuit of positive emotions from the connection to high-T objects.
CONCLUSION
Architecture has so far resisted quantification because of its complexity. The key is to identify architectural qualities that can be assigned a numerical value. The architectural temperature T was introduced here as one component of a quantitative theory of architectural design. We described how to measure T by estimating the detail, curvature, and color, and compared different architectural styles (and specific buildings) on the basis of their T value. Of course, T is only one of several qualities that contribute to architectural design. Although this paper was restricted to discussing the architectural temperature T , we indicated the direction to proceed in order to make the model broader.
The rest of the paper was devoted to arguing why a fairly high value of T is necessary in buildings. Support comes from different sources: physics, biology, and the theory of complexity. We concluded that designs with very low T don't have enough information to connect with persons. As human beings connect to their surroundings by a unique type of conceptual interaction based on a comparison with an internal scale, low-T buildings fail to connect. This purely scientific result is independently supported by experiments carried out by environmental psychologists, raising interesting questions for discussion and further research.
ACKNOWLEDGEMENT: This work is inspired by the most recent unpublished results of Christopher Alexander [10], to which this author has had access for a number of years. Alexander has developed a comprehensive theory of architecture with links to science and mathematics. Though we take a slightly different approach here, the motivation and aim of the present work are the same.
REFERENCES
- Pierre von Meiss, Elements of Architecture (London: E&FN Spon, 1991).
- Nikos A. Salingaros, "Life and Complexity in Architecture From a Thermodynamic Analogy", Physics Essays 10 (1997): March, in press.
- Stuart Kauffman, At Home in the Universe (New York: Oxford University Press, 1995).
- Herbert A. Simon, "The Architecture of Complexity", Proceedings of the American Philosophical Society 106 (1962): 467-482. [Reprinted in: Herbert A. Simon, The Sciences of the Artificial, M.I.T Press, Cambridge, Massachusetts, 1969, pp. 84-118].
- Nikos A. Salingaros, "A Scientific Basis for Creating Architectural Forms", Journal of Architectural and Planning Research (1997): (to appear).
- Robert Venturi, Complexity and Contradiction in Architecture Second Edition (New York: Museum of Modern Art, 1977).
- Robert Venturi, Denise Scott-Brown and Steven Izenour, Learning From Las Vegas (Cambridge, Massachusetts: MIT Press, 1977).
- Benoit B. Mandelbrot, The Fractal Geometry of Nature (New York: Freeman, 1983).
- Albert Mehrabian, Public Places and Private Spaces (New York: Basic Books, 1976).
- Christopher Alexander, The Nature of Order (New York: Oxford University Press, 1997). (in press).