Dave Snowden

Complexity, precision and meaning

RSS Feed

We had an excellent course in Amsterdam last week. One of those where you get challenged by the audience, always more fun that where you get passive reception. One of the attendees, Ivan Jensen, brought my attention to the following quote from Lotfi Zadeh, otherwise known as the law of incompatibility.

As complexity increases, precise statements lose their meaning, and meaningful statements lose their precision

Now I have long talked about a related concept that I call requisite ambiguity, which is a more cryptic expression of the same concept. If we are dealing with complex systems then they are non reductionist in nature. That is to say the whole is not the sum of the parts and they are non-aggregative in nature. That means as they scale the nature of language and the means by which meaning can emerge change. A complex system is a lightly constrained system, in which agents (anything which acts) co-evolve with the system itself; the system constrains the agents, but the agents constantly modify the system by the nature of their interaction with it. Managing complex system this involves a process of navigation where the map by its nature is incomplete and the topology is constantly changing. Hence the criticality of weak signal detection and the ability to quickly amplify or dampen the emergent potentiality of a system. By the time we have probabilities it may be too late, and expecting precision is fatal.

Of course all markets are complex systems, as are employees at any volume. Precision in these is a retrospective act of understanding, not a determinate or pre-condition of future decision making. We have to act before we can fully know, hence the shift from fail-safe design to safe-fail experimentation. The difficulty with measurement systems, and the whole MBA/consultancy approach to strategy with its emphasis on analysis at the expense of action is that it fails to recognise complexity and seeks precision at the cost of meaning. This is not to say that analysis is wrong, but the nature and focus of analysis is different. More on that later in the week.