Strange Objects Symposium

1

2

3

4

5

Strangeness and Agency: Mathematics Against Convention

Jason Vigneri-Beane <jcvb@splitstudio.com>

1. Introduction

The issues of strangeness, purpose, and intention in architecture are increasingly problematized by advances in both the techniques of modeling and the theorization of computing by architects. One of the major questions I think could be asked while engaging problems of architecture and computation is whether one takes the position of the experimentalist or the avant-garde. In other words, does one explore conventions in order to problematize them or does one attempt to figure the "new." Cryptography is, for me, a model for the experimental approach because it is about the reorganization of conventional materials by replacing the rules that govern convention with rules that uproot it. If mathematics replaces grammatical rules here then the organization of parts is free to permute an overall field of information that could be understood as iterative rather than fixed, information rather than meaning, one state of organization out of many. The images below demonstrate iterative understandings of architecture as a collection of states that can be contained within a single architectural complex: Romano with material-becoming-language, Scarpa with the formulation of history-language-material, Greg Lynn with formation-over-time, and Mark Rakatansky with identity-over-iterative-category. These projects always recall to me Bertrand Russell's philosophical problem of vagueness. Vagueness is, in the end, a problem of categories or, let's say, the impossibility of stable relationships among categories and objects. As Russell notes, vagueness is a problem of representation and not of objects. In other words, the flexibility of physical material will always elude techniques of representation to the extent that conventions of knowledge like classical set theory, a mixture of mathematics and language, begin to fail. In its place would then be recent lines of thought such as fuzzy logic which offer a more fine-grained, iterative, and flexible way of mathematicizing language in order allow for mixtures of categories that provide lessons for architecture - a field that is becoming less about resolution and more about the intensification of mixture.

2-4. Max Black and the Museum of Applied Logic

One of the most tectonic demonstrations of Russell's conception of vagueness and the iterative nature of objects is Max Black's Museum of Applied Logic. Here, Black proposes a hypothetical series of objects with an unfinished block of wood on one end and a finely crafted Chippendale chair on the other. In between these determined ends is an infinitely incremental series of objects that are iterating their way from the block of wood to the finished chair. The question to the relationship between language and material is whether or not one can determine which object is the line between the "chair" and an unformed collection of materials. Similar questions could be asked across different scales of architecture. 3. For example, what is the line between the building and the landscape. Or, can a surface show itself in the process of formation? 4. And then, can that process of formation produce an iterative mediator instead of the traditionally crisp line between the category "building" and category "landscape." Other questions along these lines ask at what instance a tadpole becomes a frog or, if a person is simply an accumulation of biological processes and ever-refreshing cellular systems, how can a single proper-name persist over the course of one's life? And, does the name actually persist over the stability of genetic information that can be represented mathematically?

5. Models: Material and Computational

The problems that these kinds of questions raise for a discipline like architecture, which deals with complex objects and assemblies with multiple aspects, might be engaged via research into other venues such as computing, life sciences, and physical systems. Each of these disciplines works consistently with notions of iteration, non-linearity, performance and material process Ð all contributors to the ways in which objects elude the arrest of representation. Some mathematical models of rule-based organization and flexible systems are: encryption which replaces grammatical rules with mathematical rules, cellular systems in which clusters of cells are endowed with low-tech intelligence to generate and organize other clusters of cells, particle systems in which entropy and machine intelligence mix randomness and stability over time, and distributed behavior models in which vitality and intelligent systems like flocking and schooling are simulated with simple spatial rules. Each of these models suggests the constant formation of entities that are more about logics of organization than they are simply about form.

5.5. Swarm Intelligence and Bottom-Up Organization [ SIMULATION - NOT PICTURED ]

To expand on distributed behavior models, for example, one could look at a simple program like Gnat Cloud for principles of flocking. One designs a rule-set for the simulation of a swarm of pixels by selecting rather mathematical options from various menus having to do with population, velocity of members of the population, distance maintained among neighbors, and number of neighbors to affect a component. What is important here is that one only manipulates components and local rules. Rather than designing the form of the swarm, the swarm emerges as a constantly changing formation that is stable over time. This bottom-up approach to formation places importance on rules rather than objects and offers a kind of architectural direction in the sense that the performance of systems provides a logical drive for the generation of objects that does not necessarily hinge on conventions of top-down planning. Here authorship evolves into the enactment of agency.

6. Organizational Logic over Conventions of Planning: Agent-Based Modeling

Replacing conventions of architectural organization with mathematical rules and algorithmic functions, then, offers the possibility of both the maintenance and mixture of architectural categories. These images collect a number of rehearsals toward ways in which simple algorithmic systems can organize architectural material. The bottom-up system used here is cellular automata. Each study begins with an agent-based model that uses component-agents with neighbor-dependent rules toward a generative organization of coded materials. In the top two series, for example, a cell will recognize the state of its neighbors and, based on that state, will generate a new group of cells that will, in turn, execute the same rules that produced them. This particular system grows by writing over itself and, in doing so, demonstrates the collapse of representation and production that is the hallmark of computing.

7. Agency and Mathematical Rule: Rule-Based Systems

The bottom three studies here use a system that searches cellular conditions for specific clusters. If it recognizes those clusters that are a part of its rule-set then it will grow by substituting those clusters with new clusters. What is notable about this system is that each iteration carries a collection of all possible states that the rules can produce at that moment of discreet time. Time, here, is discreet, not continuous. The condition-response action of these growth systems suggests that there is an agency to these systems that will generate organizations based on performance criteria rather than conventional decision-making processes. One designs rules, rather than objects, and rules produce diagrams that, when coded into architectural categories like, say, public/private, structure/skin, or building/landscape deliver the kinds of stealth operations that architects like Stan Allen have attributed to diagrammatic influences on architecture. There is some hidden and purposeful agency governing the formation of an object.

8. Information, Object, Category Growth out of Combination, Evolution, and Permutation of Information.

What I am proposing is a question of whether or not one can isolate deployment in order to re-mix architectural categories in a general and systematic way. The way in which the first model of cryptography produces fields of legible objects that remain unformed by conventional grammar suggests that isolating the deployment of parts from other aspects of design promotes a freedom of architectural information to form into buildings or not. Or, if a building is formed then multiplicity might be upheld by allowing mixture, entropy, openness, and evolutionary logic to exist alongside the resolution and certainty that is expected of architecture. Perhaps we could think of these kinds of operations not so much as strange objects but as strange organizations that are on their way toward forming objects. It is worth noting that the same pool of organizational information could speciate into various architectural proposals depending on the environment to which it is grafted, the material that will actualize it, or the purposefulness toward which it may or may not be arc-ing.

9. Strange Organizations

The "autonomous morphogenesis" here, to use one of Jacques Monod's criteria for strange objects, has first to do with organizational information that will form the logical, and sometimes spatial, framework for later formation of architectural objects. More specifically, this information is self-organizing to the extent that it is unpredictable but nonetheless immanent. In other words, once a system is set in motion with initial conditions like an array of cells and rules for the behavior of those cells the system will produce and reproduce the same outcome. This is one form of "invariant reproduction", to use another of Monod's criteria. In fact, there is invariance at both the level of the enactment of rules for reproduction by individual cells and the level of the overall field of generative organization that is produced by the cells over time. While it would not be accurate to claim that these small studies are truly strange objects because they require the design of rules, these studies are interested in evolutionary principles such as the accumulation of small changes over time. Specifically, the interest is in an accumulation driven by parts that, like genes, have no ability to forecast or predict the kinds of objects that will emerge from them.

10. Speciation of A Similar Pool of Organizational Information

At the same time, these particular studies attempt to claim that one could also maintain a level of abstraction and treat an environment not just as the physical site but as the modeling environment itself. Using the same pool of organizational information Form-Z, for example, sponsors one kind of development with polygon-based techniques and 3-D Studio MAX sponsors another with nurbs-over-time. This abstraction is both the interest and the problem for me when considering a systematic approach to aspects of the design process that relate to MonodÕs third and most elusive criteria, "teleonomy." These cellular systems, for example, have both a purpose and an identity. Better yet, they have a purpose that in fact produces the identity or, to use John Holland's phrase, a pattern that is stable over time. But this is not necessarily the kind of purpose that one might recognize as architecturally productive. The purpose of the cell is to recognize other cells and produce new cells based on what it sees. As an agent, it performs this act of internal logic without arc-ing toward some end, the telos of teleonomy, or responding to some kind of external factor such as a site. ItÕs strength, however, is that it produces relationships and these relationships become a kind of morphogenetic guideline that can later sponsor programmatic differences and adapt to a site.

<teleonomy . . . epigenesis . . . internal/external . . . unfolding . . . morphogenetic guidelines>

11. Purpose and Radical Uselessness

These relationships, their differences and adaptability have led me to my current position on purpose in architecture. Systems that perform some generative act of organization in a binary environment such as computing or architecture will yield coded mixtures of categories and the opposites of those categories. For example, generating an inside will also generate an outside. And generating something with purpose also generates something without purpose. This is crucial for the establishment of a rule-based freedom for architecture to be both useful and useless, to function as well as exercise itself as a medium, and to support a kind of radical uselessness. Radical uselessness is a purposelessness that emerges out of the same rules that produce purposefulness itself.

12. Organizational Logic, Adaptation to Program

For example, the agency of this organizational system is seeded to yield an epigenetic mixture of opposing categories in order to generate an overall vagueness that consists of specific opposing units of architecture. The rule set indicates that any pair of neighboring cells generates a pair of opposites. If these opposing cells are coded as "programmed" or "open" then other conditions of architectural material such as inside/outside, useful/useless, the smooth/the discrete transpose the organizational diagram into a field of architectural errancy.

13. Breakdown of Systems

Systems of architectural material such as structure and skin become mediators to both maintain and articulate the difference between the agent-based logic of organization and the accommodation of program.

I'll end here with a quote from Max Black in 1937: "The line traced by a draughtsman, no matter how accurate, is seen beneath the microscope as a kind of corrugated trench, far removed from the ideal line of pure geometry. And the "point planet" of astronomy, the "perfect gas" of thermodynamics, the "pure species" of genetics are equally remote from exact realization. Indeed the unintelligibility at the atomic or sub-atomic level of the notion of a rigidly demarcated boundary shows that such objects not merely are not but could not be encountered."