Original Dutch article      IAAA      Theory      Artificial Architecture     Remko Scha

This is an English translation of an article that was written in Dutch in May 1992, for a special issue of Wiederhall that never appeared. The original Dutch version was first published in: Remko Scha (ed.): Art, Chance and Algorithm (Department of Computational Linguistics, University of Amsterdam, 1992), and later in: Hans Konstapel, Gerard Rijntjes and Eric Vreedenburgh (eds.): De Onvermijdelijke Culturele Revolutie. (The Hague: Stichting Maatschappij en Onderneming, 1998), pp. 105-114.

Remko Scha

Institute of Artificial Art, Amsterdam

Towards an architecture of chance


For today's reader of Immanuel Kant's Kritik der Urteilskraft, it is striking that his discussion of the esthetic experience is not primarily concerned with works of art, but with natural phenomena -- its paradigm examples evoke flowers, crystals, landscapes, stormy seas and starry skies. As Jean-François Lyotard has pointed out, this is not a coincidence. It is connected with essential features of Kant's theory. For Kant, an esthetic experience is a coherent cognitive process which nevertheless does not correspond to a concept that can be articulated explicitly. Esthetic reflection is disinterested contemplation. The beautiful has the structure of the purposeful, but involves no practical purpose whatsoever.

When we are dealing with the products of human artists, these conditions for the esthetic experience are usually not fulfilled. The artist does have practical aspirations, and often the work betrays this all too clearly. In many cases, the artwork does embody an explicit idea, and a dedicated observer may in fact reverse-engineer the work and reconstruct the idea.

Lyotard: "Die großen Schauspiele der sich in Unordnung befindlichen Natur sind ein beispiel dafür, daß die menschliche Kunst niemals etwas derartiges hervorbringen kann. Denn alle menschliche Kunst ist immer nur Mimesis und letztlich suspekt, weil immer die möglichkeit besteht, daß sie mit einer absicht konzipiert worden ist und von daher ein Begriff und eine Zweckmäßigkeit mit Zweck auf ihr lastet." [1]

Process art

Esthetically motivated art thus faces a curious challenge: if it is created by humans, it will always be inferior to nature! In the course of the twentieth century, this challenge has been taken up by many artists. Several artistic traditions have developed art-generating processes of some sort -- processes which are initiated by an artist who does not try to control the final result that will emerge. Indifferently chosen readymades, chance art, écriture automatique, physical experiments, mathematical algorithms, biological processes.

LeWitt: "The artist's will is secondary to the process he initiates from idea to completion. [...] The process is mechanical and should not be tampered with. It should run its course." [2]

A new kind of mimesis: no artworks imitating the external appearance of nature, but artists imitating the blind mechanism of natural processes.

Does this answer Kant's challenge? Not immediately, because the artist still chooses the processes that are to be carried out. And the subjectivity of that choice can be painfully clear. Usually, the choice is at least partially determined by considerations which work against the esthetic experience; it always involves conventional or anti-conventional stylistic decisions which deflect the observer's attention to the art-historical positioning of the work and to its chances of success.

This problem is illustrated most clearly by mathematical chance art. The idea of mathematical randomness addresses Kant's problem in a very direct way. If the esthetic insufficiency of human art is caused by the unesthestic, practical considerations which determine people's subjective decisions, then we can try to avoid that problem by making random artworks, which have not been subjectively constructed or chosen by a human person. But, unfortunately, mathematical randomness can only operate within a completely explicit, definite framework. Most chance art is therefore conceived in terms of a rather narrow and schematic stylistic category.

Paradoxically, the extreme simplicity of chance art ends up displaying various artistic clichés with unusual clarity: the idea that an image is a grid of pixels; that it consists of a number of elements scattered over the plane; or that it is ultimately one line. Looking at a series of chance art works, it never takes long to see what they have in common, and to spot the artist's subjective decisions rather precisely.

Because in the case of chance art these decisions tend to be so conventional and unsurprising, it is immediately clear that they cannot constitute the "point" of the work. We thus end up empty-handed: the artwork has collapsed under our analytical gaze. The only thing left is a metonymical representation of the idea of "chance".

The philosophically oriented art movements of the sixties were deeply nostalgic. Most clearly in the case of the art that celebrated nothing and passivity: fixated on ideals from the past, and obsessed by their current irrelevance: paralysis. The chance art from this period partakes in that feeling: it asserts the idea of an aleatoric art production that is not controlled by humans, but it does so without any conviction. Chance art seems an absurdum, the conclusion of a train of thought that must be somehow mistaken. It is demonstrated through childishly simple little systems, and the idea of taking it seriously is not addressed.

Total chance

Nevertheless I do think that mathematical chance art is the medium that can answer Kant's challenge, and create an artificial nature which displays a complexity and a serenity that can compete with natural nature. Mathematically formulated image-generation processes can easily be combined and generalized. This makes it possible to put large numbers of chance-art ideas together into one super-chance-art-machine, which can reach a complexity that cannot be surveyed any more by individual artists.

To take a simple example: In the chance art of the sixties one often encounters programs which repeat a particular shape (usually a square or a circle) in an arbitrary, unorderly manner on different positions on the plane. Other, similar algorithms create arbitrary closed shapes by combining line segments. These two algorithms can be combined in an obvious way, so that both the shape and the position of the image elements are determined at random. Other algorithms generate a multitude of different regular patterns or regular shapes; these can also be integrated. We may thus gradually abolish choice, by avoiding the exclusion of any choice -- by affirming every choice, and by putting it on a par with all other choices inside an all-encompassing probabilistic system.

The ultimate consequence of this approach would be a computer program generating all possible images, with probability distributions that yield maximal diversity. It will not be easy to develop this program. But it is possible in principle, and we already have the technology for making significant steps in this direction.

Cage: "I was driving out to the country once with Carolyn and Earle Brown. We got to talking about Coomaraswamy's statement that the traditional function of the artist is to imitate nature in her manner of operation. This led me to the opinion that art changes because science changes - that is, changes in science give artists different understandings of how nature works." [3]

In their metonymic symbolizations of "chance", "nature" and "objectivity", the process artists of the sixties manifested a deeply felt emotion -- the desire for an art that does not originate from the whims of the individual, but from a deeper necessity. The project of a total, all-embracing chance art can now initiate the process of actually satisfying that desire -- by treating absolute randomness as the deepest necessity.

Cage: "Is man in control of nature or is he, as part of it, going along with it? [...] Not all of our past, but the parts of it we are taught, lead us to believe that we are in the driver's seat. With respect to nature. And that if we are not, life is meaningless. Well, the grand thing about the human mind is that it can turn its own tables and see meaninglessness as ultimate meaning." [4]

The post-romantic condition

The twentieth-century has engendered a multitude of artistic "isms" which is not just rich and varied, but heterogeneous and incoherent. The art of this period now seems arbitrary, ill-motivated and mannered. Individual artists may work in this tradition and yet maintain a sense of inner necessity, to the extent that they are naïve, crazy or dishonest. From a more distanced point of view, the inner necessity of the art activity is difficult to discern.

Spengler: "Man suche doch die großen Persönlichkeiten, welche die Behauptung rechtfertigen, daß es noch eine Kunst von schicksalhafter Notwendigkeit gebe. Man suche nach der selbstverständlichen und notwendigen Aufgabe, die auf sie wartet. Man gehe durch alle Ausstellungen, Konzerte, Theater und man wird nur betriebsame Macher und lärmende Narren finden, die sich darin gefallen, etwas - innerlich längst als überflüssig Empfundenes - für den Markt herzurichten."

"Was besitzen wir heute unter dem Namen 'Kunst'? Eine erlogene Musik voll von künstlichem Lärm massenhafter Instrumente, eine verlogene Malerei voll idiotischer, exotischer und Plakateffekte, eine erlogene Architektur, die auf dem Formenschatz vergangener Jahrtausende alle zehn Jahre einen neuen Stil 'begründet', in dessen zeichen jeder tut, was er will, eine erlogene Plastik, die Assyrien, Ägypten und Mexiko bestiehlt."

"Die große Ornamentik der Vergangenheit ist eine tote Sprache geworden wie Sanskrit und Kirchenlatein. Statt ihrer Symbolik zu dienen, wird ihre Mummie, ihre Hinterlassenschaft an fertigen Formen verwertet, gemengt, vollkommen anorganisch abgeändert. Jede Modernität hält Abwechslung für Entwicklung. Die Wiederbelebungen und Verschmelzungen alter Stile treten an die Stelle wirklichen Werdens."

"Und trotzdem kommt dies allein, der Geschmack von Weltleuten, als Ausdruck und Zeichen der Zeit in Betracht. Alles übrige, das demgegenüber an den alten Idealen 'festhält', ist eine bloße Angelegenheit von Provinzialen." [5]

What Spengler says here, is known today as the "postmodern" position, which declares all expressions of contemporary culture to be meaningless, and asserts that there is no way back. That the only way we can construct an image is by appropriating material from other cultures and periods, and that that's OK. That art has become theatrical, commercial and dishonest, and that that's OK.

This situation is often considered to be static, stable, and therefore definitive. Without a driving force with a particular direction, there cannot be any development. Repetition, recombination, eclecticism. The end of history.

The postmodern party-line is that that's OK too. The end of history = complete meaninglessness = Utopia. That makes a certain amount of sense, but not everybody is equally happy about that. For the Faustians among us, stasis is death -- and to acquiesce in its victory is perverse. (Spengler clearly didn't like it at all. That's where he differs from the postmodern party-line.)

"Alles, was Nietzsche von Wagner gesagt hat, gilt auch von Manet. Scheinbar eine Rückkehr zum Elementarischen, zur Natur gegenüber der Inhaltsmalerei und der absoluten Musik, bedeutet ihre Kunst ein Nachgeben vor der Barbarei der großen Städte, der beginnenden auflösung, wie sie sich im Sinnlichen in einem Gemisch von Brutalität und Raffinement äussert, einen Schritt, der notwendig der letzte sein mußte. Eine künstliche Kunst ist keiner organischen Fortentwicklung fähig. Sie bezeichnet das Ende." [6]

Unnatural art cannot be further developed in an organic way. But it can be developed in a systematic, scientific way: as artificial art. If art moves to the realm of science, its history has not yet come to an end -- because we will need at least a few more centuries to formalize and implement all our insights about perception and structure.

At the same time, the scientific approach to art makes it possible to take the postmodernist position seriously. Eventually, the development of powerful art-generation software may indeed bring about the end of current art forms. When arbitrarily large quantities of innumerably many kinds of absolutely interesting, complex, beautiful, horrible and indifferent images are automatically generated, accompanied by solid warranties of complete meaninglessness, then many of the currently practised art forms will be rendered obsolete.

If it is true that nothing means anything anymore, and all we can do is play with forms and styles, then this game may become a lot more interesting if we try to put the whole thing on the computer -- because even an unprejudiced play with existing signs seems to be beyond the power of a purely intuitive art activity. So far, the practical effect of postmodern theory has not been much more than a kind of neo-victorianism.

So that is the challenge of science to the postmodernists. If we believe that expressive art is finished, we should be prepared to move art to the domain of science. This will take art out of its isolated, marginal position. Art will be different, but better -- science and technology just happen to be the strong points of our current western culture. ["In der Generalversammlung irgendeiner Aktiengesellschaft oder unter den Ingenieuren der erstbesten Maschinenfabrik wird man mehr Intelligenz, Geschmack, Charakter und Können bemerken als in der gesamten Malerei und Musik des gegenwärtigen Europa." [7]]

Our arts evolve in the direction of verbal, explicitly articulated, scientific forms. One by one, the various art forms will be taken over and rendered superfluous by their own simulations.

Architecture and objectivity

The artistic developments mentioned above have started already, and will continue at an accelerated pace. But we do not know yet to what extent they will take place in the "art world" rather than in the cultural realms of video-clips, screen-savers, Hollywood-movies, techno-gadgets or house-parties. I noted some ideological connections with current highbrow art: the Kantian ideal of pure esthetic experience, the postmodernist awareness of cultural history, and the Hegelian trend toward concept art. But one may wonder whether such ideas will indeed determine the future developments within the highbrow art tradition. They must compete with the idea of art as communication, which is still alive as well. And there are other contingent factors: fashion, commerce, power politics, generation conflicts.

Architecture is a special case, however, since it is the realm of a deep-rooted yearning for the objective, the unpersonal, the impartial. Architecture often feels it must justify itself in the context of society at large, and transcend the conflicts of interest which exist within that society. The personal interests of the architect are suspicious. One wants pure functionality and pure esthetics.

Mondrian: "Our material environment can only posses a pure beauty (i.e., be healthy and satisfy utility in a truly direct way), if it no longer reflects the egoistic feelings of our small personality: if does not even have any lyrical expression any more, but is purely plastic." [8]

Functionalism was an attempt to develop such an objective style. An attempt that failed: the architect who focuses completely on functions that can be articulated explicitly, turns a building into a commodity which has a lot in common with a cupboard or a file cabinet. But this metaphor can also be recognized by the user of the building, and be experienced as offensive. The "objectivity of simplicity" thus gives rise to a design-language with an authoritarian look-and-feel.

"A poor person walking through Paris two hundred years ago could look at, say, Tuileries Palace, and think to himself, that's an achievement, I can aspire to that greatness and grace. A black kid today looking at, for example, the CBS Black Rock building, gets one message from such a monstrosity: Fuck you. You're not wanted. You're never allowed in here, you're not meant to know what goes on in here, and you'll never be a part of it. It is violent architecture, and it begets a violent response." [9]


Mondrian seemed to appreciate the clarity of functionalist thinking, but foresaw its problems: "For the present moment I see no possibility for arriving at a purely 'plastic' expression by only following the organism of what is to be built, focusing only on utility. Our intuition is not sufficiently developed for that, and too involved with the past. [. . .] For instance, utility often requires repetition in the way of nature (in workers housing, for instance), and in that case the architect needs the concept of plastic expression and the power to go against what the practical goal seems to indicate. Because there are always possibilities for architectural solutions which satisfy the practical goal and the esthetic appearance." [10] [Metrical repetition and symmetry are excluded by Mondrian's "laws of neo-plasticism".]

Architecture must make esthetic choices; functionalism is not a way around this. How can these choices be made, if we want them to be objectively valid? The complexity of this problem is illustrated by the failure of Mondrian's neo-plasticist solution.

Especially in Mondrian's own paintings, the "neo-plasticist laws, which determine the pure means of expression and their use," succeed marvellously in his aim of realizing a concept of visual equilibrium which does not evoke associations with the "egoistic feelings" of a "small personality". They are based on a sharp analysis of objective properties of human perception, and yield an experience which can bear comparison with the experience of nature – precisely because neo-plasticism does not imitate visible nature in any way, but merely represents properties of our own perceptual process.

The most important point of departure of neo-plasticism is, that pure expression is not concerned with the elements of an image, but with their proportions. Because the composition is constructed out of elements which are carefully chosen so that they are all perceptually equivalent, an experience at a different level emerges, where the individual identity of the elements has become irrelevant. By applying a clear and well-considered principle in a subtle and intuitive way, Mondrian created his unusually serene paintings.

But what happens when this idea is applied on a large scale in the architectural context? Then it becomes apparent that Mondrian has ignored a fundamental property of human cognition. The principles of neo-plasticism may hold very well when a "contextless" image is observed attentively, under the right conditions, by a person without preconceptions or relevant related experiences. But when the neo-plasticist principles are are applied more often, cognitive short-cuts emerge in the mind of the observer. The resulting images are not experienced any more in terms of their constitutive principles, but the image is recognized – and that is a completely different experience, which completely overthrows the neo-plasticist story. The image itself becomes an element. There are no proportions any more inside the image – there are only proportions relative to things outside of it, in the broader spatial context or in associations with past experiences. Proportions which are not controlled by the architect, and which not necessarily very harmonious. (Metrical repetition, for instance, becomes almost unavoidable now, unless we can also overthrow the structure of labour process and leisure habits.)

Total chance can thus be viewed as an attempt to realize an important ambition of neo-plasticism after all. The only way to ensure that the experience is concerned with proportions rather than with elements, is the absolute absence of style. In a chance art algorithm which may realize this ideal, the repertoire of elements and the repertoire of image structuring principles should both be as large as possible. And ever varying selections from very large as well as very small random subsets of these repertoires should be chosen to generate specific images.

In a chance architecture generated according to these principles, the visual structures are not intended to evoke a harmonious equilibrium. That may emerge at a more abstract, conceptual level – or it may not.

Postmodern architecture

Once more we see that the chance art machine is the technological elaboration of the postmodernist project. Postmodernism continues functionalism and neo-plasticism in its concern with the ideal of an objectively justifiable architecture – but it approaches this ideal in a different way. It recognizes the inevitability of style, and wants to realize objectivity nonetheless – by equating all styles, and combining them arbitrarily. A style to end all styles.

But I indicated already that today's postmodernism isn't much more than a neo-victorianism. It merely cites and juxtaposes some existing styles. Intended as a way to transcend the dilemma of stylistic choice, in practice it yields a narrow, fashionable neo-style.

Mathematical chance-architecture is the first direct answer to the demand for a genuinely style-transcending meta-style – it embodies the postmodernist idea at the right level of abstraction. In the chance-architecture project, all existing styles would have to be precisely analysed and generalised. On that basis, a mathematical definition of the space of all possible styles may be developed – a space comprising not only all existing styles, but also all their syntheses, interpolations and extrapolations. Computer programs employing that definition may then actually realize the ideal of stylistic arbitrariness.

In our first instantiations of this approach, the definition of the class of all possible styles will not be maximally precise and all-encompassing, but this can be gradually improved in the course of the next few centuries. And there is no need to postpone the application of this system until it is perfect and complete. Tentative initial versions will yield interesting results already, which can compete with the work of human designers.

Automatic design

So far I focused completely on the formal aspect of architecture, and pointed out that recent ideas about artificial art may be applied in this realm. Now I must raise the question whether that made sense, because the formal aspect of architecture cannot be treated independently of the functional aspect. If we write a computer program which designs arbitrary three-dimensional shapes, many of its outputs will be useless for most practical applications.

Within a large-scale architectural design problem, one may sometimes be able to split off certain purely formal subproblems which may be treated by chance art algorithms. But I would now like to consider the possibility of a more ambitioous set-up, which interweaves formal and functional decisions – and which leaves both kinds of decisions to a computational algorithm. I thus would like to discuss the integration of mathematical chance art with procedures for automatic design.

The automatization of design processes is one of the success-stories of Artificial Intelligence – but not because the behaviour of human designers has been simulated so well. Design programs work in a different way than humans. They use a typically "computational" approach: they find the best solution by systematically considering all possibilities.

Design turns out to be one of the areas where the dumb but accurate and systematic approach of the computer has advantages above the sloppy and conventional application of the human expert's intelligence. People find it difficult to survey a space of possibilities, if this space is defined by abstract, general rules. People conceive and interpret new situations in terms of their resemblances with situations that were experienced before. Human thinking is intrinsically conventional. A human designer therefore works with a limited repertoire of prototypes and conventions which are usually adequate, and considers, if necessary, very small variations on this repertoire.

Different procedures yield different results. A technical problem which allows different solutions that can be objectively judged as being more or less optimal, may be solved very well by a computer program which considers the different possibilities in a systematic way. In certain cases (for instance in VLSI circuit design) it is not difficult to specify the formal requirements for a design in a completely explicit way. But the space of possibilities that must be searched to find the design which best meets these specifications, may be extremely large and complex. Human persons may therefore get stuck with sub-optimal solutions, whereas more systematic computational explorations may yield better results.

Architecture is not VLSI design, but it is another area where the human designer does not function in an optimal way. An architect normally solves a new problem by invoking a standard solution and making a variation on it – a typical illustration of the conventional character of human thinking. And an architect who works differently, and tries to think through a problem without anticipating conventional solutions, is often punished for his hybris – he may very well overlook some practical issues, and his beautiful, daring, innovative building may end up with staircases which are too narrow, airconditioning which doesn't work, doors which are kept permanently closed.

It is not surprising, therefore, that automatic design technology has sometimes been applied with some degree of success in areas such as urban planning and interior design. It will take considerable effort to extend this work to the general problem of architectural design. But it is certainly worthwhile to aim for a process which constructs optimal architectural designs by means of systematic calculations based on explicitly specified functional requirements. The current performance of human architects is sufficiently bad to justify the expectation that very imperfect computer programs may be able to achieve considerable improvements.

The solutions proposed by the computer program will often not only be objectively better, but also surprising and unconventional. A provably correct and winning chess endgame played by a computer may seem bizarre and incomprehensible to a human observer. This demonstrates that the game of chess allows many possibilities which the human approach to chess does not easily discover. If we are interested in surprising and unconventional results, rather than in the repetition of human behaviour that we know already, the computer program may thus be superior.

Artificial architecture

So far, automatic design programs work within formally limited, conventional frameworks. But we can relax these constraints. We can start to develop programs which make optimal functional decisions, reasoning systematically on the basis of explicit specifications, while at the same time making their stylistic decisions in a maximally random way, by incorporating the principles of generalized chance art. Programs which generate a stylistic scheme as an arbitrary selection from the class of all stylistic possibilities, and then employ this scheme to dictate all formal decisions which do not follow from the functional specifications. In this way, the integration of automatic architectural design and chance art will give rise to artificial architecture.


[1] Jean-François Lyotard: "Die Erhabenheit ist das Unkonsumierbare. Ein Gespräch mit Christine Pries am 6.5.1988." Kunstforum International, 100 (April/May 1989), pp. 355/356.

[2] Sol LeWitt: "Sentences on Conceptual Art." Art-Language 1,1 (May 1969). Reprinted in: U. Meyer (ed.): Conceptual Art. New York: Dutton, 1972., pp. 174-175. And in: R. Kostelanetz (ed.): Essaying Essays. Alternative Forms of Exposition. New York: Out Of London Press, 1975, pp. 279-281. And in: Sol LeWitt. New York: The Museum of Modern Art, 1978, p. 168.

[3] John Cage: Preface to: "Where are we going? And what are we doing?" In: Silence. Lectures and Writings by John Cage. Middletown, Connecticut: Wesleyan University Press, 1973, p. 194.

[4] From the same preface by John Cage, pp. 194/195.

[5] Oswald Spengler: Der Untergang des Abendlandes. Umrisse einer Morphologie der Weltgeschichte. Erster Band: Gestalt und Wirklichkeit. München: C.H. Beck'schen Verlagsbuchhandlung, 1918, pp. 375/376.

[6] Spengler, p. 375.

[7] Spengler, p. 375.

[8] Piet Mondriaan: "Neo-plasticisme. De Woning - De Straat - De Stad." i10, 1,1 (January 1927). Reprinted in: Yve-Alain Bois: Arthur Lehning en Mondriaan. Hun vriendschap en correspondentie. Amsterdam: Van Gennep, 1984.

[9] Norman Mailer talking to David Frost. (Michael Tomasky: "Public Enemies". The Village Voice, March 10, 1992, p. 14.)

[10] From the same article we quoted above.