Open and Closed Design

What's the difference between open and closed? In the lab, performing an experiment to prove or disprove a hypothesis, we are working within the framework of a closed system; the original proposition governs our procedures and observations in arriving at 'yes' or 'no.' When in performing an experiment we come across something unforeseen, or are prompted by evidence to jump tracks and think about a different issue, then we are working in the framework of an open system; we move beyond 'yes' or 'no' to exploring something 'new' -- new to our own understanding.   The same contrast operates in the craftsman's studio. A potter performs a series of routines to sift and homogenize clay in order that it fire evenly in the kiln; she knows what works and what doesn't; in this closed system of practice, she is in control. But then something unforeseen in the composition of clay coming from a new source causes it to fire unevenly, no matter that she did everything right. There is a flaw, but it has a more interesting texture than those of her old pots; which suggests that she needs to rethink altogether her practices of refining clay. Now she is in the open orbit.

It might seem that the contrast between closed and open is just that of dull to creative.   But it's nothing so binary: in the lab or studio both are needed. Routines guide any kind of making, which is Lewis Mumford's point: you can't rely on the one-off, on inspiration, or on spontaneity alone; you need rules.   One reason, though necessary, these don't suffice is that a system of fixed rules can prevent improvement from occuring. Improvement requires conceiving a higher standard of rightness rather than just doing what you know will work. You have to become self-critical about what you do, but this is a tricky proposition rather than a cliche. Such improvement requires imagining an excellence which has yet to be experienced or put in practice. Not what is, what might be. The Greeks called such imaginary excellence arrete; it requires, we'd say today, an open mind.

Open and closed, once set in the context of a modern city, deviate from the idea of balance between closed guidance and open imagination. In modern cities, the closed dominates over the open, and the closed has itself become rigid in form, precluding revision or improvement. The Plan Voisin was such a closed system of design, as are all its children and grandchildren, these forests of towers which exist everywhere in pretty much the same form, and pretty much everywhere destroying informal, shifting, energizing street life.

Open and Closed Systems: Five contrasts

The grounded knowledge possessed by children in Medellin and Cabrini is intuitive, but these impressions are of a fast-changing and febrile environment as in Marx's remark on modernity that "all that is solid melts into air.” All of which suggests an absence of systematic form. Why use the phrase "open system" than simply the word "open"?

The concept of an open system was first made explicit by Norbert Weiner, the guru of cybernetics, the science of information flows and feedbacks on which modern computing software relies. But the idea of open systematicity is much older. It appeared, I think, in the idea of evolution as first propounded by Charles Darwin and then deepened by Gregor Mendel. Darwin observed that environments change in form over time; Mendel explained the genetic changes in organisms that allow them to evolve; in place of a predictable Plan conceived by a Master, the environment is an unstable process in which chance mutation plays a large part. This analogy to text-book evolution is a good way to ground modern open-systems thinking.

In general, systems analysts use five principles to draw contrasts between open and closed. First, the parts in the system which interact have a distinctive character; you cannot simply substitute one element for another. Then, simple rules can generate complex results, which is the phenomenon called "emergence." Next, known and determinate beginnings can wind up producing unforeseen or unpredictable results, which is the special meaning of "chaos" in a complex system. We can also think, less dramatically, of events which have latent consequences. Fourth, in a complex system a relatively small-scale event can trigger a massive change in the whole system; this trigger is what we call in everyday language a "tipping point", famously embodied in the flapping wings of a butterfly which through a chain of events set off a storm half way around the globe. That's of course an extreme: more normal are the changes which occur via "path dependencies," a sequence of events changing course from its original as a sequence of influences appear in its path. And finally, surprisingly, complex systems can self-organize, analyzing emerging conditions, responding to tipping points, adapting to "chaos"; such self-organization is called "auto-poeisis."

Open systems are sometimes called in sum non-linear systems, a label which path dependency makes clear: the idea here is that small events in time become magnified, to have unforeseeable consequences as with the butterfly effect. What has to be said first of all about this non-linear system is that it is not at all random, in the sense of memoryless -- history matters, only the path of historical determination is not clear or even predictable precisely in advance. In the technical sense there's prospective chaos -- and who would have thought a butterfly could become so powerful a beast? -- but also causality.

A closed system contrasts to each of these features. Mathematically, the system possesses two characteristics: additivity, which means the parts always add up to a clear sum, and homogeneity, which means parts are directly interchangeable. 2 + 2 = 4, not 3 + plus something or other, in the first case; 2x = y in the second. There is no phenomenon of emergence because the system is binary and reductive; yes/no rules. Which is to say, mathematically, that Boolean logic governs a closed system. There are tipping points in a closed system, but they never come as a surprise, they can always be calculated in advance. Additivity and homogeneity mean, further, that there is no "chaos" in the technical sense; you can produce complex structures by progressively combining simple, interchangeable elements; this is the "linear" character of a closed system. Finally, instead of autopoesis you have a different kind of self-analysis in a closed system, one ruled by equilibrium; the system treats as "noise" the feed-back of information which does not serve to maintain harmony and balance. This is why closed systems are thought of as static, and open systems as dynamic.

Every technically-minded reader of the above two paragraphs will object, "but you've left out Thomas Bayes!", the originator of probability theory, a theory whose rules about what's likely to happen mediate between the open and the closed. Bayesian logic underlies how Google and almost all other computer computational programs work, amassing vast amounts of information from your online exchanges to make predictions about what you want to buy or who you are, even though the machine doesn't know for certain. A non-linear system can include many kinds of logics in addition to analyses of probability.

The important thing now is that the five contrasts between "open" and "closed" are visceral as well as mathematical. Think of Nietzsche; in his famous work on Greek tragedy, the open system is Dionysian, the closed system is Apollonian.

With this Nietzschean distinction in mind, we can make further sense, not entirely frivolous, of some of the aspects of time in a non-linear system. Nietzsche tells us that Silenus, his Dionysian figure of excess, had lost his memory, unable to recall whenever he began drinking about all the hang-overs and vomiting which had come to him before. There are memory-less events in the dry realms of open-systems, too, like flipping a coin; the mathematician Gyorgy Markov, also an addicted gambler, spent a lifetime pondering this memory-less event, puzzling over what patterns the endless flipping of coins might produce; if there were none, he wrote, then as an addicted gambler, he was in the grip of "absolute meaninglessness."   Nietzsche's tropes of disappearance and disintegration -- the waste produced by the avid pursuit of pleasure -- are markings of Dionysian time, a time of emotional eruption and of entropy. Dionysian time appears, in the dry realm of open systems, in the phenomena of aperiodic fluctuations, that is patterns which do not repeat after a certain period, or of amplitude death, such as a pendulum gradually succumbing to gravity.

To the novelist, there would be nothing odd about the idea of an open system: it is the very essence of narrative. If after the first paragraph or two of a novel you could predict how it would end, you wouldn't bother to read it; what keeps you involved are uncertainty, surprise, and the ever-revised understanding the characters have of themselves and each other. A bad novel is linear, a good novel is non-linear.

We've so far discussed open system properties but not open system form. This is assemblage, whose form appears by drawing a contrast to bricolage. Assemblage can be likened to a chemical reaction, bricolage to a chemical mixture; in the first the elements interact because they come into contact with each other, in the second they don't interact; mixed together, they remain inert. Making an assemblage means finding out what elements will react to one another, for instance what will prompt path dependence. Will heat cause the butterfly to beat its wings faster, or will rain? Since we are within the frame of the open system, we can't predict that heat "causes" in any linear way the proverbial storm to occur as a result of the butterfly's beating wings; there are too many other contingencies and agencies. We can say that an interaction has occurred which, combined with many others, assembles the event; in time they assemble a larger whole.   Just to take the chemical analogy one step further, the assemblage isn't only composed of reactions which form stable chemical compounds; one element can react to de-stabilize or destroy other elements -- CO2 emissions into the atmosphere have this effect -- which have their own knock-on effects.

Artists have long understood the difference between assemblage and bricolage as forms. Assemblage is the idea behind collage, as done by Gris, Braque, and Picasso in the last century: scraps of newsprint, fragments of painted images, bits of photographs pasted together look different and take on a new visual significance. Bricolage is what happens on a wall covered with graffiti. The graffiti tags don't interact which one another in form; they are self-signatures the graffitist -- this is a teenager with spraycans whom a gallery is about to make rich -- puts anywhere he or she finds a wall. [The gallery converts the wall into an expensive commodity by claiming it is an assemblage, a significant whole.] In music, the chance procedures used by John Cage sometimes produce bricolage, sound events which have no relation to one another in sequence, and sometimes assemblage, as after he cast the I Ching to determine the order of sounds, so that the accumulation of sounds became the musical experience. Again, the minimalism of Phillip Glass is musical bricolage, the repetition of each fragment and it placement next to other kinds of fragments gradually composing a whole.

In the social sciences, assemblage has served as a way to get beyond post-modernism, which proclaimed the end of master narratives like Progress or like Revolution -- all to the good -- but then faltered. What instead of the grand narrative?   If there is no master narrative guiding history, there is still is history, forward in thrust even if not linear in form. Which is why sociologists like Bruno Latour, Saskia Sassen and myself want to understand assemblages.

About the Author
Richard Sennett is the Centennial Professor of Sociology at the London School of Economics and University Professor of the Humanities at New York University.
Posted on July 23rd, 2014.