Cynefin and the Chaotic domain – a (non)-review

This is the ‘review’ follow-on to the previous post ‘Cynefin and the Chaotic domain – an update‘.

Back in that post I pointed to a report by Dave Snowden that he’s posted an update on ‘Cynefin subdomains in the Chaotic’, and, as a necessary precursor to this review of that update, outlined my own understanding of the nature and role of ‘the Chaotic’ in and for enterprise-architecture practice.

Before moving to the review itself, a story that perhaps better describes the business-context for the Chaotic.

Back about ten years ago, we were doing some consultancy-work with the top executive-team of a large security-organisation. We asked what support their front-line staff had for decision-making in real-time action.

“We have rule for everything!” was the proud reply.

“What happens when you come across something that’s different, that’s new – something you don’t have a rule for?”

“We make up a new rule, of course!”

“Yes, but when do you make up that new rule?”

Some suddenly less-confident eye-flickers between the executive. “Uh… after the event, I guess.”

“So what rules do your front-line people use when they’re faced with something unique or new?”

An interesting silence…

To me, the Chaotic is anything inherently unique or new. Somewhat as with the colloquial meaning of chaos, one of the key points is that there is no immediate link to anything else, in terms of causes or comparisons. In effect, the colloquial meaning of ‘chaos’ is a context where everything seems like that, no apparent connection between anything and anything else, all happening all at once. Any conventional rule-based tactic for sensemaking and decision-making in this kind of context will usually just make things worse: instead, we need something that respects the Chaotic as the Chaotic.

Another nice example of the kind of context we need to deal here with is in ‘Points of view‘, a classic video-advert for the newspaper The Guardian:

In essence, it’s the same challenge about ‘Same and different‘ that we’re always struggling with in enterprise-architecture: most of our tools focus around some variant of ‘same’ – whether Simple, Complicated or Complex – yet the realities that we deal with always have some element of ‘different’, contextual, unknown, uncertain, Chaotic. Getting the balance right between all of those elements is hard: and to make it worse, the right balance is always changing with the ‘variety-weather‘ – kind of Chaotic-on-Chaotic.

In short, we need tools that work with the Chaotic as Chaotic, and in concert with all of the other ‘domains’. It’s what I’ve tried to do with context-space mapping, for example, or the SCAN framework, but I don’t think it’s there yet.

One of the other tools that purports to cover this space is Cynefin. Unfortunately I’ve long had a problem with this: from much experience in trying to use it in real-world enterprise-architecture practice, to me it’s always seemed ‘Complex-centric’ in much the same sense that TOGAF is still IT-centric. However, Snowden has just announced an update to Cynefin’s handling of the Chaotic: and according to the announcement-Tweet, it specifically purports to apply to enterprise-architecture and the Agile space – so I’m hoping that it should resolve my previous concerns. The review that follows focusses on this one question: does the updated Cynefin seem to address the Chaotic-domain challenges that I’ve described in the previous post and above?

[Note: What follows is my professional opinion: nothing more than that. That opinion is grounded in a heck of a lot of research and study and real-world practice, yet it is just an opinion: it does not claim to be ‘fact’, or ‘the truth’, or anything of an absolute nature. People are welcome to take it or leave it, as they choose. What I will not do is debate it with Snowden, because much painful experience indicates that there is no point whatsoever in doing so.

What follows is just an opinion. As long as you’re clear on that, let’s move on?]

—

Later: Oh, to heck with it: I can’t even be bothered to do the full review. I get so much flak and abuse from trying to sort out this mess that it’s just not worth the effort.

From looking at it again, to me the new additions just read as little more than an Aristotelian ‘Naming Of The Parts’: in effect, not just a ‘simplistic categorisation tool‘, but an over-complicated simplistic categorisation-tool – the worst of all possible worlds. Sigh.

But do go look at Snowden’s article yourself: ‘…to give birth to a dancing star‘. Make up your own mind: it’s your choice, not mine.

For what it’s worth – and it may well be worth nothing to you – my advice remains unchanged: I still think Cynefin is too’Complex-centric’ and too misleading in its handling of ‘Chaotic’-type issues such as I’ve described above, so I’d still have to recommend that it should not be used in enterprise-architectures. But that’s just my opinion: you’re welcome to do whatever you like. Just don’t say you weren’t warned, okay?

Tired…