Dave Snowden

Liminal Cynefin: the final cut?

RSS Feed

The liminal version of Cynefin has been slowly evolving over the last few months. It was getting close in August but I wasn’t happy as it didn’t properly deal with disorder or the various Cynefin dynamics. Teaching it in Sydney over the last two days allowed me to play around on the iPad while having intelligent conversations and I think its there now; but I’ve left a question mark just in case. Long term followers will know that Cynefin goes through periods of change over the years as practice and theory advance. The switch from Simple to Obvious for example was minor, while the much earlier introduction of disorder was major. The liminal version is major, and possibly one of the most significant changes I’ve made.

I’m not really ready to write it up in article form yet, so for the sake of brevity a series of bullet points:

    • There are only two liminal areas as obvious to chaotic is a cliff and obvious to complicated an overlay based on human perception which, unlike all the other boundaries in Cynefin, does not represent a phase shift.
    • The liminal area in complex, bordered with complicated is in effect a domain of experimentation, or pilots, prototypes and iterative experiments.  Scrum, Lean start up and the like all fit here as they start with an idea of what will work and seek to make it so while being prepared to vary or even abandon it.  A state of suspension, of holding options open for as long as possible, characterises this space.
    • The complex domain, excluding the liminal area, has higher levels of ambiguity with early stage exploration, unarticulated need emergence, parallel safe to fail experiments and so on.  There is a qualitative difference in terms of ambiguity and stability of hypotheses between the main domain and its liminal state.
    • My old shallow dive into chaos, now becomes a liminal area in the chaos domain, or possibly an intrusion from the complex domain.  Here all effective constraints are abandoned to allow novelty to emerge; agents are not allowed any connection in tools like MassSense and so on.  But it is contained, it is a bubble and drawn as such with a closure unlike its liminal sister.

  • My major insight this week was to realise that the liminal line bisects disorder and creates an authentic legitimate domain of transition, while the rest is an inauthentic state of confusion.  In the former we know what we don’t know, in the later we don’t know what we don’t know.
  • With this treatment of disorder I can now handle the dynamics which are a critical part of Cynefin. The shallow dive from entrained expertise (red) is now potentially a hazardous change, while the grazing dynamic (purple) is a state of near continuous liminality. (see illustration)
  • It also means that the rest of Chaos is red, a place in which you try and control the exit, but it will always be temporary in nature.

Now there is more to come here, but for the moment this at least provides a basic note for the record, and for comment. I will now abandon the domain frameworks in their current three by three forms as the liminal version works better, but there are aspects of those that I need to pull out into supporting frameworks – all part of a composite body of material on a complexity based approach to strategy.

The banner image represents a liminal time of the year!
La noche de Walpurguis by Mariano Barbasán Lagueruela