A rant on new wine on old wineskins (again)

I had a conversation a short while ago with someone who had been using a “complex systems method”. I was told that the originator of this method advocated small experiments which is goodness. However it appeared that her instructions were to do one intervention at a time to see if each one worked before progressing to the next. Now this is arrant nonsense. Its another example of someone grabbing the language of complexity but fitting it into the strait jacket of a linear approach. The essence of the shift to safe fail experiments is that you have multiple experiments in play. This will be the case anyway, a complex system is not a closed one that can be experimented with in this way. In practice it may be a combination of experiments with in situ action that works. The point is that you simply don’t know. Critically all experiments must be tested for coherence, and must be monitorable and also safe-to-fail; but other than that let a thousand flowers bloom.

I often think some of this sort of nonsense is the consultant trying to control the process rather than facilitating it.