Which among these three human artifacts is best described as a system: a hunting net, a shoe, or a gasoline powered motor? I think many of us are inclined to answer the third as being the most system-like. Why?
When one is asked to give prototypical examples of some concept, it is natural enough that we choose those examples with which we are familiar. When we think about a concept we often do so with these prototypical examples in mind. This entails a kind of ethnocentric coloring to our thinking about concepts, including that of systems. While this bias deserves acknowledgment, perhaps it is a bias without serious consequence. I put it forward here to dismiss it from immediate consideration. Yet before I do, it is interesting to contemplate whether our intuitions would lead us to consider the artifacts of one kind of society to be more system-like than the artifacts of another kind of society. For example, are the typical artifacts of modern techno-culture more system-like than those of a simple horticultural or hunter-and-gatherer culture? If so, why?
Returning to the original question, what differences between hunting nets, shoes, and motors account for our inclination to describe these human artifacts as being a system or system-like entity? Or to put it contra-positively, what is it about systems that make hunting nets, shoes, and motors more or less systems-like? Wikipedia (article entry on Systems
) gives a concise definition of system as "a set of interacting or interdependent entities forming an integrated whole." This superbly general definition (of which I will have more to say later), will serve as a baseline for our comparison.
The Gasoline Powered Motor
First let us consider the motor, our chosen paragon of an engineered system. From our (general or specialized) knowledge of gasoline powered motors we know that a gasoline powered motor works as a smoothly integrated functional unit composed of many finely designed part each of which fits neatly into an designated place (physically and functionally) in the designed whole (if only while the motor is in working condition).
Many of the various components of a motor are clearly recognizable as such; see, there they are! the epistemological cut is given to us directly through our perceptual apparatus because the decomposition is largely a "natural" one, one reflecting what one might call first-order "fractures in the fabric of being", those gross non-uniformities more or less directly accessible to our perception; in this case these non-uniformities are artifacts of a design and manufacturing process emphasizing modularity.
This does not mean that combinations of more basic components do not comprise subsystem components in their own right, (such as the entire electrical system of an automobile) or that one could not if one chose, go deeper down to even the sub-atomic level in describing a motor as a system. Clearly the granularity of one's decomposition will depend upon one's purposes and one's knowledge (assuming now that only veridical decompositions are considered).
Note though that the interaction of the parts occurs when the motor is in operation. The operation of a motor cannot occur except in some larger situation of which it is part. A motor requires fuel, and an oxidizer to run, and there must have been some mechanism by which the motor was started (e.g. a human using an ignition switch). Indeed, the situation in which a motor finds itself will or will not support a variety of different motor behaviors. Our default notion of the behavior of a motor is set against a variety of background conditions (of which we may or may not be aware), the change in any of which will change the operation of that motor. For example, if we try to operate the motor in extremes of heat or cold, or immersed in water, or in a vacuum, or with kool-aid instead of petrol, we can expect that the motor will act different in each, and that the parts will interact in (possibly subtle) different ways. If a motor is bathed in a large pool of sulfuric acid, the system loses its structure, sub-components of sub-components, sub-systems of sub-systems disassociate mixing with the pool, the relations between parts dissipate as the parts from which they are made dissolve, increasing the total entropy of the total system. An acid dissolves a system from the bottom up.
Yet, our concept of a motor may require no explicit background conditions in which to "operate". Imagine deep empty space, and sailing across your field of view, a motor in heated operation. Of course, we know that a typical gasoline powered motor could not operate in such conditions, but we are capable of imagining them as entities independent of any particular context (though I'm afraid that I see no way in which to imagine a motor in a way which does not set it against some kind of background, even one relatively empty of content.) That we can imagine what is not actual should serve as a simple lesson, one neatly conveyed in this recent XKCD cartoon:
What about the shoe? It has clearly recognizable parts, and these parts surely interact with each other in various ways, though the parts may not fit neatly with our usual classifications (sole, tongue, laces, etc) with which we describe them.
When a person walks in a shoe (or pair of shoes), a whole range of forces will be applied to the shoe depending on the person's specific activity. The forces will be transferred through the body of the shoe and the foot in various ways that depend both on the structure of the foot and on the design of the shoe (such as the cushioning in the sole). When a shoe heats up or cools down, the various parts of the shoe expand or contract in different ways depending on the substances from which they are made, applying or relieving pressure on other parts of the shoe with which each component has contact.
Yet, for all this, of the three a shoe feels intuitively least like a system. That does not mean that it is not a system. In fact, I suspect that a knowledgeable shoe engineer would be able to give a nuanced description of shoe performance in terms of its parts and the relationships between its parts-- in short, as a system.
I might acknowledge, at least in the abstract, that a shoe is a kind of system, but fail to consider a shoe to be a good example of a system, i.e. not a candidate prototypical example of system-ness. There are at least two possible reaons for this. First, it could be that something about my ordinary or naive concept of a shoe lacks some quality or set of qualities that would make it seem system-like to me. Perhaps the way I regard shoes does not include a notion of how the various parts of the shoe interact. I might not even think about shoes as having parts, except when I have to (although the same may be said about my thinking about motors). Second, it could be that shoes simply do not have a rich enough collection of easily individuated parts and enough interaction between those parts to make them good candidates for a prototypical system. So while a shoe is a system, it may somehow be less system-like, or at least, less impressive, than other examples of systems which come to mind.
The Hunting Net
I chose the example of a hunting net for a variety of reasons. First, I wanted to use an example less common to the every-day experience of a lot of people, including myself, but an artifact common to the lives of many others. I've also been reading a copy of Collin Turnbull's The Forest People I picked up at a thrift-store. In the book Turnbull spends some time describing the Mbuti's use of hunting nets. Finally, a hunting net seemed suggestive of a continuous field, but one in which various forces acting on parts or regions of the net pull and distort other parts of the net.
What are the parts of a net? This is a somewhat difficult question to answer. A net will have warp and weft yarns, knots or loops where the warp and weft yarns intersect, spaces in between the yarns through which things like air or water may pass through. Also, the net may have different shapes (i.e. different constraints on the shapes the net may take). One might talk about the inside and outside of a net, depending perhaps on the current configuration of the net (e.g. when something is caught inside). But one might want also to define regions of the net, or include particular lengths of yarn, or perhaps the fibers of the yarn in one's decomposition or description. But it is worth noting that what is distinctive about being a net is not as easily capturable by examining the relations between individual atoms (into molecules), or molecules (into fibers), or fibers (into yarns), as it is at the level of yarns woven into net.
Perhaps it is this uniformity in a net's design that resists a simple functional decompostion. The various parts of a motor and of a shoe all have distinct and complementary functional roles, but to my untutored eye, one part of a net seems to have much the same function as another any other. Perhaps a more nuanced notion of the construction and use of nets would permit such a decomposition.
Instead of looking at functional roles of parts, perhaps it is more productive to begin thinking about the pattern of interaction between points in physical structure of the net, as in a field. (however we define those parts, vague or otherwise). If someone were interested in simulating the cast of a net using some suitable computational model of the net, then that person might represent the net at any given moment as a point in a some suitably large state-space. As I've written before, a state-space is transformable into a family of system of parts by designating which different states belong to which types. Hence, we might define particular regions of the net when it is suitable, or define different regions in the state space when the net is some kind of state, such as being tangled, or torn.
We confronted with an object there are two issues. One: whether the object is a system, or not a system. Two: whether or not we count that object or that type of object as a good example of a system. That an object is a good example of a system or not depends on both the actual properties of the object, and on our concept of that object (type). Both of these involve the complexity of the object, our knowledge about those object-types, and our particular interest in entertaining the object (as system). There generally seems to be a natural scale at which a decomposition is most appropriate, but while some objects facilitate decomposition by their physical nature and/or designed functional modularity, some objects resist an obvious decomposition into easily discernible parts. But any state space may be given a suitable, if arbitrary, classification of sets of states into types of situations. In all cases, much interaction between parts, the behavior of the system, depends on the situation in which it is found. Normal or default operation of the system is set against a variety of background conditions. Some background conditions destroy the relations between parts characterizing a system.
A Last Note
Finally, a note on the meaning of integrated in the definition of system given at the top. Some would define integration as bringing together into a harmonious cooperative whole. Such a definition is probably fit enough for the kinds of objects we've discussed, particularly the motor. Yet, I think the requirement of cooperation and harmony may be too burdensome. Clearly there are system-like objects which are not harmonious and do not function as cooperative wholes. The weather systems, for example. There exists an alternate more forgiving definition of integration as unifying into a whole, which suffices to relate the information that we are regarding the object as composed of parts which taken together with the relationships between them make up that whole.