Two Sides of the Thinking Coin

As is often the case, it is the neglect of fundamental principles that is at the root of poor performance. Sports coaches often stress the return to fundamentals—the “blocking and tackling”—as the key to improving team performance. This is a recognition of the powerful effect that adherence to sound fundamental principles and practices has on performance.

Systems engineers are as prone to ignore or gloss over the “fundamentals” as any athlete. They become caught up in the nuances of “advanced” practices and techniques and fail to recognize the importance of the foundational principles of the systems practice.

Systems thinking is a paradigm
Dictionary.com defines paradigm as a “framework containing the basic assumptions, ways of thinking, and methodology that are commonly accepted by members of a scientific community.” The community of systems disciplines (systems thinking, systems dynamics, systems engineering etc.) shares in a variant of the systems thinking paradigm. The paradigm focuses on the concept of systems as distinguished from sets. While sets are simply collections of elements, systems are collections of elements organized in particular ways for the accomplishment of some objective or objectives. Systems thinking concerns itself with systems.

Most definitions of a “system” contain some variant of the International Council on Systems Engineering’s definition: “A construct or collection of different elements that together produce results not obtainable by the elements alone. The organization of the elements of a system is called a construct” because the elements exist in relationship to each other. It is through this construct, or set of relationships, that a system can produce results not obtainable by the elements acting separately. Systems thinking is concerned with the three aspects of systems—the elements, the construct and the results.

The importance of systems thinking in systems engineering
The idea of a system construct is central to systems thinking. Where systems thinking is used—as it is in systems engineering—to design system solutions to system problems, the ability to see the system construct is critical. It is through the construct or relationships that the system produces its results, so an inability or failure to see the construct makes it impossible to understand how the results are produced.

In understanding the problems to be addressed, it is critical to understand their genesis. They arise from the interactions produced by the system construct that creates them. To understand the problems, we must be able to see and understand that construct.

The same is true of the candidate solutions under consideration. These solutions are systems whose results must interact with the problem systems in ways that will “solve” the problems. The solution solves the problem by producing interactions that will alleviate its detrimental effects.

It is only by understanding the system constructs of both problem-generating and solution-offering systems that solution results can be matched with problems in a meaningful and effective way. This is accomplished using sound systems thinking.

Two aspects of systems thinking
Our Western scientific paradigm has its origins in the Enlightenment—specifically in the thinking of René Descartes and others. This thinking calls for the understanding of systems to be secured through a process of reducing it to its simpler parts. The idea is that the simpler parts can be more easily grasped, and the understanding of the larger system can be the result of an aggregation of the results of the simpler understandings. The process, reductionist in nature, is what we know as “analytic” thinking. (Note: The word analytic itself has roots in the Greek for “loosen” or “break apart.”) Analytic thinking is the foundation of the scientific method flowing from the Enlightenment.

This process of reductionist thinking followed by the addition of the elemental level understandings into a “system” level whole has value and works as advertised to some extent. Particularly where the systems being analyzed are constructed in ways that the actions of the system elements have linear cause and effect relationships to each other and the system outcomes are deterministic, analysis is helpful in reaching system understanding.

Such deterministic, linear cause and effect constructs are most commonly found in complicated systems. While they can have large numbers of elements, they are characterized by linearity and deterministic results. For that reason, analytic thinking can cope with understanding all but the most complicated.

However, the world is increasingly trending away from complicated systems and toward complexity. In complex systems, not only are there apt to be a large number of elements (detail complexity) and an intricate series of changes (dynamic complexity), but also the results produced by the system are the product of or are contributed to by emergent behavior. The results of the relationships in a complex construct are not typically linear or determinant, making causation hard to track at the element level. The emergent behavior cannot be predicted by adding together the individual elemental results. The only way to understand or predict the results from a complex system is through understanding the system as a whole, beyond the sum of its parts. This is achieved not through analytic thinking, but through synthetic thinking.

Synthetic (the word has its roots in the Greek for “put together”) thinking focuses on seeing from the whole rather than from the parts. This kind of thinking is the key to understanding complex systems. Without it, the understanding of emergent results and complex relationships is lost.

Systems thinking involves both kinds of thinking
In its most powerful form, systems thinking combines the relative strengths of both the analytic and synthetic approaches. Analytic thinking allows us to understand the properties of the elements as they stand alone while synthetic thinking leads to the understanding of the elements in combination, particularly in complex combinations.

Constructing and selecting solutions requires the understanding and predicting of their results. The same is true for dimensioning and understanding problems. Without the application of both kinds of thinking, the picture is incomplete and inaccurate. The predictions of the results of applying solution constructs to specific problems are correspondingly inaccurate. This jeopardizes the success of the solution application in the problem space.

So where do we go wrong?
Most systems engineers would agree with this reasoning. They would accept the premises that would require analytic and synthetic thinking processes. Usually, analytic thinking is the more obvious and easily applied of the two. This is most likely because it is the thinking that is embedded in our Enlightenment scientific paradigm. We are taught to break apart problems and solutions in order to understand them. So applying analytic thinking comes almost reflexively.

Remembering to think synthetically is a bigger challenge. It isn’t “natural” for us. The systems level viewpoint is often harder to get. Therefore, in many instances synthetic thinking goes begging. Outside of our observance of an analytic paradigm, how does this happen?

Choosing processes, methods and tools
We have a natural tendency to choose analytic tools. Our tools often focus on one class of elements without setting them in the system context. We have requirements tools, component architecture tools, risk tools and others which deal with one kind or another of the system elements. While they fill an analytic niche—understanding that particular class of elements—they do not show us the element in relation to the other classes. They do not show us the element in the greater context of the whole. This denies us the synthetic view.

Other tools represent more than one class of elements but confine themselves to a particular subset of system information. Drawing tools fall into this class. The information about the system design is bounded by the four corners of the drawing. There are a number of tools that advertise their ability to provide a set of views as a “system model.” The views may be underpinned by a data dictionary of objects to place on the drawings, but each drawing is effectively maintained independently. In doing so, the tools are not producing a system model that will support a truly synthetic look at the system.

This is because a true system model must fully integrate all of the system elements into a single construct where they can be seen and evaluated together. Where this does not happen, the fundamental benefits of synthetic thinking are limited or denied by the broken relationships. The recent RFP issued by the Object Management Group for a “SysML v.2” recognizes this shortcoming in calling for the proposal of a “SysML v2 metamodel” that would frame the construction of a system model to tie together the “views” of a system design. Without such a meta-model/model, the synthetic view is missing.

Another way our tools can impede the synthetic thinking process is by over-simplifying the ways in which we can represent the realities of our solution. In the name of accessibility, some tools limit the richness and nuance of expression that can be used to construct the models that we use to grasp the reality of the problems and solutions we face.

Barring unusual enrichment or deprivation, a two-year-old child will have a vocabulary of 900 to 1000 words. Contrast that with the average 12-year-old who will have a vocabulary approaching 50,000 words as well as a much more nuanced working knowledge of language construction. The 12-year-old can use the larger, nuanced language knowledge to make more accurate, understandable expressions of her ideas than the 2-year-old who is constrained by her limited working vocabulary and syntax. By limiting the vocabulary and syntax, tools can decrease our ability to communicate our ideas around problems and solutions. As complexity rises, this limitation and its effects are exacerbated.

We should choose tools that do not limit the variety and richness of our model construction or expression. Having more ways to relate the elements in our models is good. Having more views to share our models with is good.

What’s a systems engineer to do?
How do we benefit from attention to these fundamentals? The answer can be simply put—remember and practice them.

Remember:

1. Effective system engineering requires systems thinking—about problems and solutions. This is because systems engineers must be able to accurately predict and influence how their system solutions will perform in the context of their problems. This requires a systems view.
2. Getting a systems view requires being able to see the system elements in full relation to each other and to the system context. This means our system models must be fully integrated.

Practice:

1. Tool choices that offer an underlying meta-model that provides a framework for a system model that is fully integrated,
2. Tool choices that provide a broad vocabulary and syntax for constructing the model and sharing it with others, and
3. Don’t settle for less—doing so compromises our effectiveness.