c.1 Conclusion

The starting point for this research was an unmistakable fascination for the process of self-organization. At first sight, spontaneous ordering is something that occurs in nature in systems with a countless number of components, whereas the man-made built environment is essentially simple. It consists of finished plans and its shapes are constructed from blueprints that arrange the world through lines. But as complexity theory of cities has shown, this is a matter of perspective.

Not only when zooming out to the whole of a city through time, but also when zooming in to the most direct surroundings of people, patterns become visible that a classical explanation of order falling into chaos fails to explain. This can be vividly seen when looking at stuff, that crystallizes around places and activities in one-to-one configurations.  The aim of this abductive research is to find an explanation for this observation by exploring stuff from a complex perspective. Can the spontaneous organization of stuff be a process of self-organization?

This question immediately raises a problem, as stuff cannot create connections by itself. All of its patterns have to be explained through the interaction with conscious people with cognitive abilities. Here we follow the theory of Portugali explaining why cities display complex behavior; they are not solely the environment acted upon, but are also the medium of interaction and the input of cognitive processes (2011, 2016). This doubly complex model acts as the basic premise of this research, that aims to translate it to a smaller scale.

Fig c.1: The city as a doubly complex system as the basic premise of the research, which is translated to a smaller scale level.

Doing so, however, leads to another obstacle. Cities are large-scale collectives and exist over hundreds of years, which makes that the urban fabric can never be individually controlled. Translating this theory to the scale of stuff, means an answer has to be given to why, also on a scale where every item can in fact be arranged, it is still the accumulation of local interactions that, at least partly, creates its form. This is possible, of course, because of the directness of this human-stuff interaction. Whereas decisions in a city are large scale plans that take place once in a while, but are considered altogether over a prolonged period of time, the decisions to move stuff are small, occur in a high frequency and very directly follow perception.

Projecting a theory of self-organization on stuff thus has two requirements. The first is the assumption that there exists a bidirectional link between action and perception. The theory of affordances, originating from psychology and recently introduced in interaction design and architecture, presents a framework for this interaction, stating that the perception of our environment happens through action possibilities or what objects afford us to do. Affordances themselves are constantly ad-hoc created by the configuration of stuff, which is again what we perceive; a circular causality implying a process of self-organization. The second is an interaction between action and cognition. This is explored by action identification theory, also from psychology, that proposes a two-way interplay between the cognitive construct of action (what one thinks one is doing) and action itself. This similarly indicates a self-organizing process, in which the action identity works as an order parameter filtering incoming information and structuring the activity.

When these interactions take place, the organization of stuff is the result of this doubly complex process, in which both external information, (affordances) and internal information (action identities) serve as an order parameter creating structure over time. These requirements are not always met. When no activities take place, nothing happens. If stuff is allocated to a specific place beforehand and is displaced according to the drawings, there is no self-organization either. A high feedback-sensitivity between action and environment occurs especially when people are immersed in a space and create it from within.

If so, the above theory can very directly describe the forms and patterns that arise between artifacts. All stuff cells include a central point, the primary working field. This place, in front of the user and in the middle of all other stuff, acts as the order parameter of its global form. It is also the place of the most complete one-to-one assemblage of artifacts and thus is the core of what the stuff configuration affords to do and where the ‘actual’ activity takes place. This primary working field follows the same complex dynamics as action identities and affordances when it continually makes phase transitions to higher or lower panarchical levels, thereby organizing the stuff around it. The figure below gives an overview of stuff as a complex system.

Fig c.2: The constant configuration of stuff explained as a doubly complex system which acts on the two levels of function (action possibilities, affordances) and cognition (action identities). These result in a self-organized pattern of artifacts that follows systemic phase transitions over different panarchic levels.

c.2 Outcomes

In the abductive theory of method new models are explored via a process of theory generation and theory development, resulting in various explanations to be reflected upon. The most important criterion for this reflection is that of consilience, or explanatory breadth, stating that the best theory is the one that can explain the greatest range of facts (Haig, 2005). As the development of the above theory was a constant process of iteration, this research has built upon one explanation, which differs from the classical paradigm.  This model, that describes the configurations of stuff as a doubly complex system can describe our observations where a simple order-chaos alternation fails, and is therefore, following this criterion, the best explanation. The limitations of this research are in the method; abduction searches the theory that is currently most complete, but gives no proof of law. This means that the conclusions of this research are per definition open for new suggestions and improvement.

This research was not initiated because of the direct relevance of a societal problem. Instead, its starting point was the opportunity around a scientific problem; the idea that findings from other disciplines could shed a new light on a phenomenon that is all around us. The outcome is therefore not a list of principles that work as a direct guidance for design, but a theoretical model that provides a different perspective to observe what already is.

But most importantly, in this process of constructing theory, the self-organization of stuff is made explicit. As shown in the last chapter we now have a basic vocabulary, a pattern library and a clearer idea about the impact of control parameters to properly discuss what we see. This again results in more precise observations. Instead of something non-existing or mythical (even scary), the constant displacement of stuff can become a topic for discussion in the process of design. A general framework, that can from now on be built upon, no longer excuses architects to see stuff as something personal that is impossible to relate to, neither as something to take complete control over. The self-organization of stuff is an integral part of every building, and can, with the ways of understanding given above, be thoughtfully considered.

c.3 Implications

How people interact with their environment is a fundamental topic for architecture. Self-organization gives a thoroughly different perspective than the classical explanation, which is why a discussion on the implications of this view can be endless. In this paragraph three common and related design discussions are reflected upon with the results of the research, in a first attempt to formulate what this theory implies for architecture. Consecutively, these are the trend of minimalism and decluttering, the discussion on design control and the question on whether we, when stuff self-organizes, need design at all.

c.3.1 Self-organization vs minimalism: creativity and innovation

The self-organization of stuff requires a constant interaction with activity. Because of the circular causality between the two, this reasoning can be turned around; activity requires the self-organization of stuff. In this light we can reflect on the trend of decluttering, or minimalism, that since around five years has gained immense popularity, to be noticed in the abundance of organization gurus and the thousands of youtube videos on how to adopt the lifestyle (Bijlo, 2016; Bottelier, 2016; Brodesser-Akner, 2016). Lately, however, an increasing number of people has backed out, stating that minimalism makes their life calm first, then boring (Urist, 2017). Messiness again gains popularity, such as pleaded for by economist Tim Harford. Values like creativity, acceptance and resilience are integral to disorder as it can help generate ideas, he states (Harford, 2016).

Fig c.3: Internet meme satirizing architects and their workspaces.

This research concludes that actual creation is not messiness but self-organization, which indeed needs random ideas to both develop and innovate (and not grow in one rigid pattern). To engage in creative activities, stuff is needed to both fulfill the required affordances and to generate ideas, which explains why extreme minimalism can lead to passive behavior. The creative process also needs order to specifically pick artifacts as soon as the cognitive idea on what to do arises, which is why also owning an overwhelming amount of stuff can eventually lead to passive behavior. An alternating pattern between the two states, both in space and throughout time, is according to the theory most vivid, and luckily also most natural. Preplanned activities, in which the internal information on what to do is unambiguous, are more efficient when artifacts are ordered, but even here slightly messy workplaces have led to great inventions. The possibility for adhocism, as Charles Jencks and Nathan Silver describe the improvised assemblage of stuff so beautifully (1972), is sacrificed when everything is strictly allocated a place.

c.3.2 Self-organization vs control: an external brain

The fact that the emergence of action is a cognitive complex process, in which internal and external information alternate, implies that stuff is an active and essential part of our thinking. Besides generating ideas in the process, stuff is also an ad-hoc constructed part of our memory; through the external information we create, we can communicate with our later selves. Stuff serves as a reminder and structures activities (by the stuff in a stuff cell), daily life (by the stuff cells in a house) and even identity (by the whole of stuff in someones life). Stuff generates its own language, in which the artifacts are words, but their formations are the sentences that carry meaning.

This second result makes it possible to react to practices of architectural design in which all elements are perfectly harmonized, such as buildings that are strongly committed to a certain style. In buildings that are mainly created to be experienced by the senses, such as a cathedral, design control is a justified approach. In houses, however, a strict top-down control is very risky as it can, according to the theory, even impair cognition. The need for people to use a space as a mind map, or a large drawing, to continuously add to and reflect upon, makes adaptability essential. In a completely controlled environment without the ability to do so, people cannot extend their memory via their surroundings, and thus miss the possibility to structure their thinking in this way.

c.3.3 Self-organization vs the question: do we need design at all?

One anecdote that keeps coming back in literature on the topic is that of MIT’s Building 20 (Brandon, 1994; Harford, 2016). When the institute found itself in need for extra space in the middle of the Second World War, the young architect Don Winston was given the task to design a 18,500 square meters barrack as quick as he could. The first designs were ready in the evening and the construction phase that followed was finished only a couple of weeks later. The building soon became infamous for its uncomfortability; in the summer it was too hot, in the winter it was too cold, all was dusty and wayfinding was extremely difficult. It was cheap, weird and incredibly ugly.

Fig c.4: MIT’s Building 20 in Cambridge, Massachussets. A simple wooden frame made up the main construction.

Nevertheless, Building 20, was loved by all of its occupants. They found in it the perfect place to experiment, not despite but because of its sloppiness (Harford, 2016). Made for temporary use in wartime, the building miraculously survived until 1998. In the years in between it had seen the invention of the first atomic clock; Spacewar, the first arcade game; the theories on linguistics by Noam Chomsky and those on cognitive science by Jerry Letvinn; and even an improvised anechoic chamber, in which John Cage first imagined 4’33”. By the time of its demolition, Building 20 had been home to nine Nobel Prize winners.

Building 20 is a legendary example of a place that empowers self-organization of stuff. But it also reveals one of the most fundamental questions that this research raises for architecture. Self-organization is a phenomenon that seems to occur anyway, regardless of what we design. Do we need design at all?

Reflecting on this question from the perspective of the research gives three possible answers. First of all, Building 20 is a building, and not an empty lot. Although primitively, it provides for basic needs at least to a point where stuff can take care of the rest. Where it was leaking, buckets caught the dripping water, but at least it did not rain. When it was too cold, people wore thick sweaters, but at least it did not freeze. The building, just by being a building, created the most essential conditions as a starting point where stuff cells could further build upon.

Secondly, building 20 was famous for its awkwardness. The counterintuitive layout of and irrational corridor and office numbering, let people regularly bump into each other or enter the wrong room. Because of all the different disciplines sharing the building (simply everyone that did not fit elsewhere on the campus), scientists from all different fields met, leading to the most surprising collaborations and innovations. This seemingly occured through a lack of ‘design’, but can also be seen as a feature of the building. The design was great in generating random external input for people to stumble upon, leading to ideas from unexpected perspectives.

Fig c.5: People bumping into each other in the long corridors of Building 20. Many disciplines crossed each other’s paths.

Fig c.6: The improvised Rad Lab in Building 20 during the Second World War.

Thirdly, the building allowed for extreme reconfigurability. Power cables and water pipes were exposed and easy to reach, and besides were visibly old and cheap. Nobody cared what was done to the place; people painted the rooms, used doors as tables and even expanded their territory into adjacent offices by completely pulling down walls (MIT Institute Archives, 1998; Harford, 2016). Whereas scientists in other buildings had to wait for months to get permission for their experimental setups, in Building 20 these could be improvised in hours. In other words, almost all possible affordances could be ad-hoc constructed, which not only resulted in many experiments and activities, but also gave the occupants the possibility to give their room and the building as a whole its own identity (MIT Instutite Archives, 1998). More than pure ‘flexibility’, when alterations are reset as easily as they are made, the process was that of constant rebuilding, a course of iteration. Although Building 20 might have been erected in a few weeks in 1943, the actual construction took place in the 55 years that followed.

Do we need design at all? Building 20 shows we do, for exactly those three reasons. The encouragement of the self-organization of stuff happens when the extremes of the weather are, at least to some extent, evened out, when unexpected inspiration is at hand and, most importantly, when this can be converted to the creation of affordances by the adaption of one’s surroundings. Although not in final or polished form, by providing these three things, Building 20 was surely a manifestation of design.

c.4 Towards design

Understanding the self-organization of stuff gives us much opportunities for spatial design. By using a complexity vocabulary, we can describe the life cycles of stuff and other layers in a building in a dynamic way; by identifying its patterns, we can use stuff as a tool of analysis and make attempts to simulate it in our designs; and by understanding control parameters we can link it to conditions that can be provided by architectural design. When we want to stimulate this process, which can be for a number of reasons, we can take in mind what is said above; create the right conditions, add some randomness and make the design transformable to everything imaginable.

As shortly noticed in the introduction, there is another interesting avenue to explore. The self-organization of stuff is an ever-present driving force, that creates form. In fact, Spoerri used it to paint. As architecture is undeniably about order, could we not extrapolate this potency? Is it, with the knowledge we now have, possible to let architecture emerge from within, in time, as a constant game between the found and the looked for, the parts and the system, the system and its surroundings?

The next explorative section of this website, directions for design,  explores the possibilities of the theory in the design process.