Emergence is the magic of a system. As the entities of a system are brought together, a new function emerges as a result of combination of the function of the entities and the functional interactions among the entities. Recall that a system is a set of entities and their relationships, whose functionality is greater than the sum of the individual entities. This second phrase focuses purely on emergence. “Greater functionality” is delivered through emergence.

Nothing “emerges” in the domain of the form of the system. When the form of chunk A and chunk B are brought together, nothing more happens in the form domain than the sum A plus B. Properties of the aggregation of the form are relatively easy to calculate. The mass of the system is just the mass of chunk A and chunk B. Form is “linear.”

However, in the functional domain, the sum of A and B is much more interesting, is much more complicated, and has nothing to do with linearity. When the function of chunk A interacts with the function of chunk B, almost anything can happen! It is exactly this property of ­emergence that gives systems their power.

Striving to understand and predict emergence, and the power it brings to systems, is the primary goal of system thinking.

Another way to understand the importance of emergence is to think what might happen if the anticipated emergence does not occur. This can happen in two ways: The anticipated desirable emergence can fail to occur, or undesirable unanticipated emergence can occur; see Table 2.1. Both are bad situations.

Consider some potential system failures in our examples. In the human circulatory system, a connection from the heart may partially clog, raising blood pressure in other parts of the system; thus an unanticipated undesirable emergence occurs, impacting the entire system. In Team X, a team member may develop good requirements but not communicate them effectively, and the team consequently makes a design that causes the company to lose a sale; thus an anticipated desirable emergence fails to occur.

Gridlock on a four-lane urban expressway is a classic example of system failure. Every car is doing exactly what its function calls for, transporting people. The roadway is performing its function, supporting car travel. The drivers are staying safely behind each other and within their lanes. Yet the anticipated desirable performance of cars traveling rapidly down the road fails to materialize.

A well-documented case of unanticipated undesirable emergence is shown in Figure 2.12. An Airbus A320 tried to land in a crosswind with the upwind wing held low. On the slippery runway, the wheel brakes were not effective, so after touch-down the pilot tried to apply the engine-based thrust reverser, but it did not operate. What happened? As a safety measure, the software system was designed not to deploy the thrust reverser until the plane had “landed,” which was signaled by weight compressing both landing gear assemblies. But because one wing was held low, the landing gear on one side was not compressed. Everything worked exactly as planned that day, but the result was a system failure—an emergency.

Trying to understand and anticipate such system failures is also a goal of system thinking.

As suggested by Box 2.8, the last task in system thinking is to predict emergence. It is hard to predict a priori what will emerge from the combinations of the functions of the various entities of a system. The anticipated desirable function may emerge (system success) or may fail to emerge, or something unanticipated and undesirable may emerge (system failure).

We have seen examples of anticipated function emerging. In the amplifier circuit, amplifying voltage plus setting gain created amplification. In Team X, the members working effectively and communicating created a good design. But if the systems thinkers did not already know about this emergence, how would they predict it?

There are three ways to predict emergence. One is to have done it in the past. This is prediction based on precedent. We look for identical or very similar solutions in our experience and implement them with at most small changes. We build the pendulum mechanism in grandfather clocks because our grandfathers built it that way. When Team X was put together, there was experience that suggested that this group of people would form an effective team.

Another way to predict emergence is to do experiments. We simply try putting together the entities with the proposed relations to see what emerges. This can range from tinkering to very highly structured prototyping. You could explore the output of the Op Amp by building one, applying an input voltage, and monitoring the output. Spiral development is a form of experiment in which some of the system is first built to check emergence before the rest of the system is built (in later spirals).

The third way to predict emergence is modeling. If the function of the entities and the ­functional interaction can be modeled, then it may be possible to predict emergence from a model. One example of spectacular success in modeling occurred in the development of integrated circuits. IC’s with billions of gates are now routinely manufactured and produce the correct emergent properties. How? The fundamental element is a transistor, which can be modeled simply, and from there up it is all mathematics. Our amplifier circuit can be modeled with a few lines of ­algebra if one knows the constitutive relations for the Op Amp and resistors, and Kirchhoff’s voltage and current laws.

What do you do if you need to predict emergence for systems that are without precedent, ­cannot be experimented on, and cannot be reliably modeled? Welcome to the question at the crux of system thinking! Such issues arise routinely in many domains, including new product ­development. In these situations, we are left to reason about what will emerge. This reasoning may be informed partially by precedent (observing results in similar but not identical systems) and partially by experiments and incomplete modeling, but the projection about emergence ultimately depends on human judgment.

Notice that the emergence of function from a system depends on the function of the entities and their functional interactions. The form enables the function of the entities, and the formal relationships are instrumental in functional interactions. This implies that both form and formal relationships (structure) are important to consider in predicting emergence. To see this more clearly, let’s look at the simple systems shown in Figure 2.13:

• An electrical low pass filter, consisting of a single resistor and a capacitor. Place the resistor between the input voltage and output, and the capacitor between the output and ground, and high frequencies are attenuated. Change the pattern of connectivity by switching the two components, and low frequencies are attenuated.

• A mechanical lever, consisting of a bar and a fulcrum. Put the fulcrum near the end farther from the operator, and you get the desired emergence of “magnify force.” Change the location, putting the fulcrum closer to the operator, and this desired ­emergence disappears.

• A simple software segment, with a conditional statement and a simple ­computation. Put the IF statement first, and the “a = 100” statement is executed only if the condition is true. Change the sequence, placing the a = 100 statement first, and it is always executed.

The formal relationships are critical to the emergence: the pattern of connection of the electrical components, the location of the fulcrum, and the sequence of the software instructions. The formal relationships are important in guiding a certain specific functional interaction that leads to a specific system-level emergence.

In summary:

• Emergence occurs when the function of the entities and their functional interaction combine to produce a new functionality, which is more than the “sum of the parts.”

• System success and system failure often hinge on emergence.

• Emergence can be predicted a priori by relying on precedent, experimentation, and modeling. For unprecedented systems for which experimentation and modeling are not easy, humans must reason about emergence based on available information.

• Emergence depends on the function of the entities enabled by form, and on the ­functional relationships enabled by the formal relationships.

• It is the property of emergence that gives systems their power and also creates the challenges in understanding and predicting them.