“Complicated” vs. “Complex”

Someone asked for comment, I offered it. The background: A gentleman who makes a living “reducing complexity” for IT systems keeps running into some of us on Twitter who study complexity. You can imagine the entertaining exchanges, which led to his posting these observations on his LinkedIn group.

The ask:

There is a group of complexity aficionados that criticize my use of the word “complexity.” In general, these are folks who are influenced by the Cynefin framework that considers complexity and complicated to be different attributes of a system. I reject their use of the word “complicated” to describe what I call “complex” for three reasons.

First, complicated describes a state of mind of the observer, not the observed. To me, my car is complicated. To my mechanic, it is not complicated. Yet the car hasn’t changed.

Second, complicated has no obvious relationship to simple. In my mind, complexity and simplicity must be closely related in the same way that heat and cold are related. Cold is the absence of heat. Simplicity is the absence of complexity. This relationship is absent in the Cynefin-like understandings.

Third, my use of the words complex and complexity conform to standard usage of the words.

That my position. What do you think?

My answer:

It is difficult to know where to start in responding. There are not only entire graduate courses (http://www.thegreatcourses.com/tgc/courses/course_detail.aspx?cid=5181) on the topic of complexity, there are actual institutions involved in related research. (http://www.santafe.edu) Rather than attempt to capture all of the works from Ashby (1962) to Waldrop (1992) – all of which predate the Cynefin framework – I will try to address the main errors I see in your thesis. (I decided to include a select bibliography in case anyone who comes across this thread is curious enough to follow up.)

First, unless you refer to your physician as a “biology aficionado,” your language regarding scholars strikes me as somewhat odd. In my view, someone who pursues higher education is something other than an aficionado. More importantly, it appears to you that the people who contest your view above emerge from some curious cult of Cynefin practitioners who cling to definitions to support our unique worldview. This is probably the chief error in our conversations – believing that Cynefin somehow provides the intellectual framework for our understanding of complexity. Cynefin is a sense-making framework that leverages natural systems understanding – it stands on the shoulders of giants. (My bibliography below contains only one reference to my friend and mentor David Snowden.)

To your three points:

1) To your mechanic, your car, unless it is 20 years old or so, is actually quite complicated thanks in part to pervasive embedded computing. This is why he uses diagnostic technologies to understand its current state. Complicated systems are the playing field for domain experts, but that does not mean these experts consider their systems as simple.

2) Complicated has a remarkably obvious relationship to simple. Cause and effect are related in these ordered systems – and system behavior is predictable as a result. In your mind, complexity and simplicity are related like heat and cold. Standard usage aside, complexity is not some spice added to a complicated system – that mucks up the gears when it reaches a certain threshold. (I’m reminded that Aquinas famously used the heat and cold analogy to discuss how evil is the absence of god.) But analogies aren’t proofs – generations of systems science are not discarded because of a tidy analogy. There actually are relationships among systems in Cynefin, but not as you cast it above.

3) This is what you’ve repeated in past conversations. It’s how most people think of complex – and you are comforted because you conform to ‘standard usage.’ Pardon me, but I’m reminded the common approach to medical problems 100 years ago involved leeches. In my view, “standard usage” is an insufficient understanding for those who seek to advance a practice.

Overall, it is noble work to emphasize simplicity for IT systems. I was honored to support the National Military Command Center years back, and can recall the massive tangle in trying to understand all the systems that were introduced into that facility over time. Different functional requirements, different budgets, different departments, all installing systems over a few decades with no obvious design authority. But among that mess, no new systems emerged. Nothing became sentient and engaged in semi-autonomous or autonomous behavior. Systems did not exhibit any behavior that exceeded their original design. The facility, a system of systems, was incredibly complicated, but demonstrated none of the characteristics of a complex adaptive system. One can pursue this noble work without introducing terminology that confuses more than it illuminates.