What makes something a "SYSTEM"?
The notion of the "system" has evolved to help with the general problem of complexity. The "systems approach" reduces the apparent complexity of a situation by building models that can be tested against reality and against the needs of the problem solver. (Note that the complexity of the actual situation stays the same. In reducing the "apparent complexity" a systems model throws away information. The hope is that by eliminating unimportant details, we can focus on the aspects of the situation that matter the most to us in terms of our analysis.)
Various definitions of "system" have been proposed. All of these include at least a collection of individual objects and/or processes that communicate with one another in such a way as to produce a network of connections. That is, each item in the system is connected to at least one other item in a way that leaves the whole collection fully interconnected.
All definitions of "system" also include an agreement that the network which makes up the system is embedded within an environment. The environment includes everything that is not inside the system ... a definition that sounds rather vague, but which serves the purpose of isolating the system for study while accepting the fact that it cannot actually be removed from outside influences.
Systems theory is the attempt to describe, define and explain the basic idea of a "system". Systems thinking is the attempt to use systems models as part of a method for problem solving, design and management.
The "Parts" Model
One of the earliest definitions looked at a system as a collection of parts which are interconnected with, or related to, one another and which also relate to the environment which surrounds the system. In the picture below, the circles and rectangles represent the parts, the solid lines represent the relationships among the parts, and the arrows show the system's interaction with its environment.
To say that the elements of a system are interconnected implies that if one part changes, then at least one other part must change, too. Naturally, as soon as that second part changes, some other part must then change ... and so on. This is somewhat like the effect of touching a bowl of Jello - a single touch results in a long period of jiggling motion.
Because systems interact with their environments, they are constantly being "touched" from the outside. This means that most systems are constantly changing, and, because these changes take time, a system cannot be described as having one particular shape. This property makes systems useful for studying the kinds of situations that we usually refer to as events, or processes.
The idea of a system is well illustrated by the artistic device called a "mobile." The parts of this system are represented in the illustration below as "fishes." The relationships are established by the bars, which maintain a horizontal spacing among the fish, and the pieces of string, which keep the fish at certain vertical depths.
Notice that the strings and bars
- connect every fish with every other fish,
- allow the fish to move around quite a bit, yet confine them to a certain area and keep the mobile from falling apart.
If any one fish moves, at least one other fish will react by moving, too. Air is this system's environment, and the smallest breeze will keep the mobile in constant motion.
The following quotation by Stephen Littlejohn provides a more formal definition of "system":
From the simplest perspective, a system can be said to consist of four things.
- The first is objects. The objects are the parts, elements, or variables of the system. These objects may be physical or abstract or both, depending on the nature of the system.
- Second, a system consists of attributes, or the qualities or properties of the system and its objects.
- Third, a system must possess internal relationships among its objects. This characteristic is a crucial defining quality of systems. A relationship among objects implies a mutual effect (interdependence) and constraint.
- Fourth, systems also possess an environment. They do not exist in a vacuum but are affected by their surroundings. - Littlejohn, 41
Clearly, the "fish" mobile meets these requirements.
The collection of parts approach does a very good job of making clear the composition of the system and its boundaries. However, because systems can include many objects, each with many relationships, this approach is less useful in terms of practical problem solving. One of the most popular ways of modeling more complicated systems is shown next.
The "Flows" Model
Consider the familiar "homework grading system" that exists in many school settings. The elements of the system are students, teachers, assignments, homework, and grades. The relationships are various communications between teacher and student. The environment includes parents who are concerned about the student's work load and grades, and the institution which employs the teacher.
In this model the elements sometimes originate a communication ... for example, here the teacher is the source of homework assignments and the student is the source of completed assignments. Sometimes the elements receive a communication ... in this case the teacher receives the student's homework while the student student receives a grade.
Two of the elements are special. The "teacher" element originates the flow by creating the assignment. Such elements are known as "sources". The "student" element terminates the flow by receiving the grade. Such elements are known as "sinks." The grade may be passed on to the student's parents, but these are in the environment and not considered as part of the system's flow. Similarly, the institution may influence the teacher's choice of assignments, but this, too, is in the environment.
The "flows" model is quite abstract ... which means that you see diagrams instead of the actual things ... but it has the advantage of being able to measure the flow of communication, and it shows precisely how changes in one flow affect changes in others. In this example you can see that the yellow flows are creating items, while the blue flows are notifications that an item now exists. You can also see that "assignment", "homework", and "grade" involve an ongoing interaction between the student and the teacher.
The elements in this system are labeled ... "homework", "grade", "teacher" ... but they are not very well defined. In this approach to systems thinking, these are known as "black boxes" -- their inputs and outputs are specified, but their internal workings are unknown. In fact, each of the black boxes might be a system itself ... in which case, it, too, could have a diagram that showed how it worked.
The ability of the "flows" method to clarify the results of system interactions makes it quite useful as a management and design tool. However, it also has an upper limit of complexity. The next example illustrates that limit.
(Note: You may want to take a look at the tutorial on how to read and make systems diagrams.)
A Whirlpool Example
Consider a stream of water flowing between two banks. As with all streams and rivers, it has a direction of flow ... from higher ground to lower.
Now think about a whirlpool of water within that stream. This whirlpool can be defined as a collection of water droplets, air molecules, 
debris and other materials. Each of these elements is connected to the others by means of physical forces that bind them together into a swirling mass that is distinct from its environment ... namely, the stream, its banks and bottom, and the atmosphere.
Parts
One way to look at this is to take a very close view that sees the individual particles of water, air bubbles and other individual objects that make up the whirlpool. 
In theory we can track the motion of every particle. If we could establish the relationships among all of the particles, and also their relationships to the environment of the whirlpool, we could produce the "collection of parts" model as described earlier.
In practice there are two difficulties with this approach. First, there are so many particles involved that we simply cannot track them individually. Scientists deal with this by applying statistical methods, thus reducing the individual particles to mathematical expressions that represent their tendencies to behave in particular ways. The result is quite useful ... but it does not provide the insight into individuality that the "parts" model requires.
Second, the environment of the whirlpool is not materially different from the whirlpool itself. Both are composed of particles of water, bubbles of air, and so on. The identity of the whirlpool is established by its swirling behavior ... not by its physical components. In fact, molecules of water may enter the whirlpool, flow around in it for awhile, and then exit by rejoining the stream.
(This last concern is especially important. Many systems have a changing roster of elements. An "educational system", for example, constantly graduates students and matriculates new ones. The teachers and administrators change. The textbooks change, as does the furniture and even the buildings. Yet, we envision it as one system that persists over time.)
So while this approach gives us a fair approximation of how the whirlpool looks like, it does not say much about how it maintains its shape.
Flows
A wider view would include the stream and show the whirlpool as a material flow.
Water (with bubbles, debris, and do on) is flowing downstream ... suddenly some of it spins around and flows upstream for awhile. There is also a spiral of material downward towards the stream bed. From this perspective we might diagram the flow as shown here.
The model includes a source, a sink, and four processes. The upstream supply from which the material comes is the source, and the downstream sink represents the outflow that continues down the stream.
The process items do things to this flow of material. The Attractor selects some material and brings it into the whirlpool ... material not selected simply continues downstream. The Spin Cycle pushes material upstream ... the TIME element represents the time it takes for the material to get upstream ... at which point it again encounters the attractor, which may or may not pull it back into the whirlpool. The Downdraft element sucks material down ... after which it continues to spin. Some material may be pulled far enough down that it escapes the whirlpool and flows out via the downstream sink.
This model describes how the whirlpool works. One of its advantages is that the black boxes can hold statistical data. For example, we would not need to know exactly which water molecules are selected by the Attractor ... we only need to know the percentage. That means that this model could be programmed into a computer and used to simulate the behavior of the whirlpool.
However ... gaining the ability to model the whirlpool in the abstract, loses the sense of what the real whirlpool is actually like. In practice the tension between these two approaches is seldom resolved. Either the model shows good detail, but cannot handle great complexity. Or, it can deal with complexity, but the detail gives way to statistical estimates.
This inability to completely model reality is one of the shortcomings of the systems approach. On the other hand, the biggest and best computer models ... those that model the weather, for example ... have the same shortcoming. It may be that the human brain itself has a limit to the amount of detailed complexity that it can handle.
Some Problems With Systems
Arguably, system thinking is the best method yet invented for dealing with highly complicated processes. But it is far from perfect. There are a number of ways that a systems analysis can go wrong, and people who use systems thinking have to pay careful attention to these.
A number of books already provide examples of the limitations of systems, and rather than add more examples here, I would refer you to:
- The Art of Problem Solving: Accompanied by Ackoff's Fables (Russell L. Ackoff)
- General Principles of System Design (Gerald M. Weinberg, Daniela Weinberg)
- The Systems Bible: The Beginner's Guide to Systems Large and Small (John Gall)
Here are three general situations that can cause problems for systems thinkers.
Boundaries
While the parts model does a better job of describing the system as we might encounter it in the world, the flows model does a better job of explaining how the system actually works. But each in its own way points out the problem that confronts all systems models: it can be very difficult to separate the system from the environment.
This is called the "boundary" problem. At one point, there is the system. At another point, there is the environment. In between, there is almost always a boundary of some sort that is both in the system and outside of it.
In the homework example suppose that the student sends the grade report home to his or her parents. At what point does the grade actually leave the system? When the student puts it in an envelope, when it goes into the mail, when the parents open the envelope? In this particular example it probably doesn't matter much ... but in many situations, the boundaries become extremely important.
Boundaries occur at points where communications enter and/or leave the system, and their careful definition is a required part of the systems modeling process.
Changing Real World Conditions
The system is intended to model a situation as it exists in the world. But the world is constantly changing, and if the systems design process does not take account of this, the final system model may no longer match reality.
Consequently, the modeling process needs to include frequent reviews of the data, and if the data changes, the model must be updated to fit.
A second consequence is that systems models tend to have short expiration dates. A model that was useful in solving a problem last year may be outmoded when a similar problem reappears this year.
The important insight is that while the finished model is the result of the modeling process, it is the process that matters. Every problem requires the development of a new systems model because the thinking that goes into making the model is the key to solving the problem.
Wishful Thinking
The results that you get from a systems analysis depend entirely on how you define the system in the first place. Because of this, some philosophers argue that systems theory is really just a way of rearranging what we already know, and that it never discovers any new facts. This may be true ... and if you find this debate interesting, do a web search on "systems theory philosophy" and begin reading. It's a fascinating subject.
But it is not relevant to this discussion ... which is about using systems as a means of thinking about problems. In fact it is exactly the ability to rearrange what we already know that makes systems thinking so powerful. In many cases the "problem" only exists because we haven't looked at all the facts, or because we haven't looked at the facts from the right point of view, or because we stubbornly refuse to admit that there might be a second way of doing something. Sometimes the act of defining the system is all we need to "solve" the problem ... which in fact simply dissolves when we see it from another angle.
However ... to create a system model you have to make choices about what to include in the system and what to exclude to the environment. And not surprisingly, there is a human tendency to include things that support the outcome you want and to exclude things that run contrary to your preferences. If you're not careful, the result can be a system that models your fantasies rather than reality.
Successful systems practitioners learn very quickly that wishful thinking is not a very good method. However much you want a particular outcome, or fear a particular outcome, you simply must exclude those feelings from the model building part of the process. The right approach is to build accurate models that are flexible enough to accommodate multiple outcomes ... and then to criticize and appreciate those outcomes.
For the same basic reason, systems models cannot be compared to one another. A systems model is the result of a careful crafting of elements, attributes and relationships designed to fit certain data to a particular situation. Two different people will model the same data differently depending on their particular problems and preferences. (You can argue about the processes ... about how each person analyzed the situation ... but the two models are simply the results of these processes and are not comparable.)
It's worth repeating that the finished model is only the result of the modeling process and not worth much outside of the particular situation that's being modeled. It's the insights that the model enables that matter. (Which is why it's called systems thinking.)
Are Systems Real?
I have been careful to describe systems as human constructions. That is, a system is something that I make. Some people argue that systems are in fact real ... that the world is composed of actual systems ... and that the models we make simply reflect this reality.
This could be true ... or not. For the purposes of these tutorials, the question is irrelevant. If what we are actually doing is making models of the real world, then well and good. If, on the other hand, what we are doing is building abstract models that permit us to have insights that we would otherwise miss, then also well and good.
(As noted above, if you are interested in the philosophical questions, a web search on the phrase "systems theory philosophy" is a good place to start.)
Systems Examples
The following examples illustrate the contents of this tutorial:
Example 1. College Life
Example 2. Fossil Fuels