At the moment, I am reading and thinking a lot about complexity and how it could be applied to development and enrich the Systems Dynamics Analysis I am using in my work. Today, I read an article by David J. Snowden and Mary E. Boone titled “A Leader’s Framework for Decision Making” and published in the Harvard Business Review back in November 2007. Snowden and Boone added a box to their article in which they describe the main characteristics of complex systems. I found this to be a very comprehensive and yet understandable description an that’s why I want to share it here.
Here you go:
- It [a complex system] involves large numbers of interacting elements.
- The interactions are nonlinear, and minor changes can produce disproportionately major consequences.
- The system is dynamic, the whole is greater than the sum of its parts, and solutions can’t be imposed; rather, they arise from the circumstances. This is frequently referred to as emergence.
Moreover, Snowden and Boon differentiate between two types of complex systems. In the first type, the individual actors or ‘agents’ in the system strictly follow predefined, simple rules, such as birds flying in a flock or ants in an ant colony. In the second type, however, the individual agents are not animals but humans and, hence, follow their own reasoning according to the relevant context and situation.
Consider the following ways in which humans are distinct from other animals:
- They have multiple identities and can fluidly switch between them without conscious thought. (For example, a person can be a respected member of the community as well as a terrorist.)
- They make decisions based on past patterns of success and failure, rather than on logical, definable rules.
- They can, in certain circumstances, purposefully change the systems in which they operate to equilibrium states (think of a Six Sigma project) in order to create predictable outcomes.
So where does this lead us in our everyday work? Snowden and Boone also offer a number of tools to manage complex situation out of which I want to pick two that I find are relevant for the work in development projects:
- Open up the discussion. Complex contexts require more interactive communication than any of the other domains. Large group methods (LGMs), for instance, are efficient approaches to initiating democratic, interactive, multidirectional discussion sessions. Here, people generate innovative ideas that help leaders with development and execution of complex decisions and strategies. (…)
The first point clearly points out that participation still is a very important part of every development project that really wants to make a difference. In the end we have to be aware that it is not us that is changing the system, but we are merely working to enable the system to move itself towards a more favorable state (who defines whether this state is more favorable remains another point to discuss and influences a lot whether the system is actually moving in that direction).
The second point is important to recognize that we always have to look for things that work or try to start small pilots and see whether they work and amplify them. This is essentially the recognition that change to a system happens from within a system.
I will continue blogging about complexity, many things are going on in that field. So stay tuned.
Pingback: Linking to Complexity « Shawn Cunningham's Weblog
Your article contains a lot of good points. I am also reading and thinking a lot about complexity and how it could be applied to the management of our systems. I have a blog over complexity but in German (http://www.anchor.ch/wordpress).
I think that a real complex system is on a high level of organisation. Complexity isn’t aequivalent with chaos. A real complex system is the brain I think. It has a high degree of structure which isn’t static but re-make itself again and again. This may lead to a sort of path dependence, and it is a good idea to anticipate windows where a path break becomes possible. I think the more complex a system is the deeper the groov of its path dependence and the smaller the windows of path breaks.
Also I don’t think there are complex and non-complex (complicated?) systems. For me this distinction isn’t helpful. There are only systems which are more complex than others. Complexity is a relative term.
I found a paper with the title “20 Definitions of Complexity”. Some definitions are very specialised, others less significant, less meanigful. But the paper shows that there is no stringent definition of complexity. Nobody knows exactly what complexity is but everyone uses the word….
In social contexts complexity could also have a subjective dimension.I don’t mean that some situations only seem us to be complex. There is no demon who could objectively measure the level of complexity. The situation is such complex we recognize it.
Dear Peter, thanks for your comment which I appreciate very much, especially because they come from a fellow Swissman. You brought up some very important points on some of which I will certainly blog again in the future. Let me comment some of them right away:
On the level of complexity: you say that there are no un-complex systems, it’s just the level of complexity that’s different. Well, I think that is a matter of definition and I don’t want to go into semantics here. I agree that there are certainly different degrees of complexity. I do, however, think that a differentiation between simple, complicated, complex, and chaotic is indeed helpful when approaching a specific problem. Each of the field requires different approaches. I found the Cynefin Framework very helpful to define the different spheres of simple, complicated, complex, and chaotic. The framework was developed by Dave Snowden – one of the authors of the paper I cited in this post. You can find the Wikipedia entry here and a very informative YouTube video by Dave Snowden himself here.
On measuring complexity: I just started reading Melanie Mitchell’s book ‘Complexity – A Guided Tour’. I’m in the first part, still, where she exactly brings up the question on how to measure complexity. What’s more complex, she asks, a yeast or a human being? I hope to share some thoughts on her proposed definitions soon.
I’d be happy to get a copy of the paper ’20 Definitions of Complexity’ you mention. Is it available for download? Could you share the link with me?
It’s simply downloadable on web.mit.edu/esd.83/www/notebook/20ViewsComplexity.PDF
I know the Cyefin Framework, thank you for the hint. It seems to me rather a simplification. The difference between complex and complicated is a little bit outdated and a trial to be fancy by hook or by crook. But you are right: is a matter of definition.
Hi Peter. Thanks for the link, I’ll certainly have a look at the document. Now to your argument that the Cynefin Framework is a simplification: yes, certainly. But that’s how we work, isn’t it. We try to build models that are on the one hand manageable by our brains being limited in the capability to grasp complex interrelations and on the other hand are close enough to the reality to still give some reflection on it.
I don’t think the differentiation between simple, complicated, complex, and chaotic is outdated. For me it still makes a lot of sense, especially to decide what set of tools we can use to solve problems. The most obvious examples are light switches (for simple problems), jet engines (for complicated problems), and the economy (for complex problems). To put it bluntly: you wouldn’t start working on the economy with a screwdriver. You do need different tools to work in these different areas.
But to be honest, I also still have some problems really finding the right grasp on how to separate complex and chaotic. They seem to be very much related.
Hi Peter: thanks for sharing the MIT paper. It look very interesting.
In my opinion the discussion about complex systems is a discussion about complex models of ( parts) of reality. There are no complex systems in reality, systems are constructs of the mind. useful no doubt, but constructs still. The management literature tends to forget this. I think, because of their Anglo-Saxon orientation, MIT in particular.
So the question here is really about the need to build more complex systems of reality in order to understand what is going on. Presumably simpler models are not helpful enough in solving puzzles. But on the other hand there is a limit in what we can comprehend with regard to the number of variables we built into our models.
Considering we are talking about social systems ( at least I am) I find the thinking of Niklas Luhmann about social systems very helpful because by following the principles of autopoiesis ( look at the system you construct about reality as a system that has as its goal reproduction, regard the system as a closed system, although it accepts inputs from its environment) you reduce the complexity of the model of say an organisation.
Hi Ton, thanks for your comment. Interesting point you are bringing up. I agree with you that what we work with are only models of the reality. Models are not supposed to be complex, since we design them to be simple, still reflecting some of the most important caracteristics of the complex reality. Besides the models that give us insights into the micro-sphere of the systems, we also need to understand the macro-dynamics of complex systems. When we combine this knowledge, we are better able to achieve the changes we want to see in those systems.
Well Marcus, as we can perceive the reality only by models – everytime our brain makes a model of the external world – the world would never seem complex to us if there aren’t complex models as you said. As complex the reality may be we have no possibility to recognize this complexity as we have no possibility to recognize the reality itself.
Well, you are going pretty philosophical here. I think we should stay practical. Of course we need models so we can break down the reality into peaces we can swallow and discuss. Still we need to be aware that we are talking about models. The art is to build the model in a way that it retains the most important characteristics of the real world. This is really how many sciences work. Once we understand the model, we can go back and test our solutions in the reality and then we see how accurate we were. Based on the experiences, we can then improve the solutions.