Dave Snowden

Falcons and white feathered pigeons

RSS Feed

Cynefin.pngMany years ago I read Max Boist’s Knowledge Assets and came across his I-Space model which inspired an early version of the Cynefin framework. I have always acknowledged the debt, but at no stage laid any claim to Boisot’s original model. This is basic academic integrity; acknowledge your sources but take responsibility for what follows. Integrity also requires its practitioners to not claim status by association; a form of plagiarism that should probably evoke sympathy as much as condemnation.

Since the Cynefin framework achieved some fame I have seen various responses which can be roughly grouped as follows: (i) use and development of the model, with proper acknowledgement of source, (ii) various uses of the model to domains with which I have varying levels of knowledge and (iii) what I will call the vindication response. The first two of these are legitimate and a stimulus to thinking, equivalent to my work with Boisot. The vindication approach is one where someone seizes on the model as vindicating something they have believed for a long time, but have not been able to secure the attention they feel they deserve. All too often they have not even read the material, just seized on some of the language and and key phrases with which they resonate, and then use the model to vindicate their own ideas; bootstrapping credibility. This is the second form of plagiarism that I reference above and I will confess I generally get irritated before I realise that sympathy or indifference is probably a better response. That said I always try my best to engage with the individual concerned.

Observers of the twitterverse will by now have realised that I am referencing recent disagreements with Tom Graves, self described as an Enterprise architect, business anarchist and confusionist. I hadn’t paid much attention to the most esoteric of these descriptors but over the last month ago I have gained more appreciation! Tom is also into water and pendulum dowsing, “earth-mysteries” and personal development based on the nordic concept of Wyrd; an interesting combination of interests which has some relevance to what follows.

Origins of Cynefin in Systems Architecture

The original reason I engaged with Tom is that he appeared to understand that a systems architecture is not just about IT systems. This contrasts with my first encounter with the subject, back in IBM days where the founding father was generally considered to be J A Zachman and his IBM Systems Journal paper A framework for information systems architecture is still a classic. There are two interesting links here. The first fully refereed article on Cynefin was also in the IBM Systems Journal, and the referees made several references back to Zachman. Secondly there was an extended discourse between the Cynefin group in IBM and various systems architectural initiatives. IBM was then (in a CEO sponsored initiative) engaged in a research programme to take the ideas of a CEO (possibly with the involvement of his/her board but I got the sense that was optional) and translate them into IT systems with minimal human intervention. That discourse which went on over a couple of years had a profound impact on the emerging Cynefin framework. The “Probe-Sense-Respond” nomenclature along with the catastropic fold between simple and chaos all come from that period. Since that time the application of the Cynefin framework to business design has continued, but there is also a significant technical interface in our development of both tools and methods.

In the development of SenseMaker® Ken took an approach based on complexity in which the majority of our spend has been on an object orientated architecture in which people can also be considered as objects. We gave it a first showing recently to a major systems integrator and the response was gratifying. Our goal to create an architectural framework that avoided the linear and taxonomic approach that are all to common appears to have succeeded and survived its first third party review. We aren’t in a position to give details yet other than in the context of a strict NDA, but I raise it now to demonstrate that since leaving IBM the whole idea of a design/architectural approach has been basic to our theory and practice. The major differences I had with others in IBM was over this need to keep human networks continuously engaged in interaction with technology objects. Its an approach which fits the philosophy of many a systems architect, but much of the practice has in effect become an extension of Business Process reengineering. The taxonomic structure of Zachman’s developed work illustrates the point well, as does the Gartner Group practice of attempting to create order from diversity (may the Lord have mercy on their souls because surely any sane user will not). We have taken a theory based approach, and that theory is based on the science of systems, not our observations of cases or any particular ideological commitment (other than the naturalistic of science).

… and now we come to the tragic bit

Screen shot 2010-06-06 at 09.43.55.png

Note, the links here are to pdf files not to Tom’s blog as he has started to censor contributions there and in consequence I can’t trust him not to change the content. So I have captured the versions as of this morning and loaded them as files.

I’d been following Tom for some time and knew he was interested in Cynefin. However a post in Feb 2010 provoked a response from me. He produced a modified version of the Cynefin framework (see top model in the box to the right) and stated:

“I’ve been concerned that there’s been very little attention paid to the role of the Chaotic domain. So that’s the theme I want to tackle here: how may we reclaim the Chaotic, to make Cynefin more complete?

(I’d better say upfront that there’ll be a fair amount here that Dave and others may disagree with, sometimes quite vehemently – and that’s okay, because this is definitely a ‘work in progress’, and probably with gaping holes in the reasoning in places. I need that critique if this is going to work in practice. In no way do I consider that any of the other work in Cynefin is somehow ‘wrong’ – particularly not the work that Dave and others have been doing in the Complex space, which I regard as crucially important in business and elsewhere. All I’m suggesting here is that perhaps we need to approach the Chaotic domain with the same degree of discipline as we do with the others – and not simply ‘run away’ to the Simple or the Complex as soon as we hit the Chaotic, which is about all that standard Cynefin offers at the moment.)”

Now he was correct that there was much for me to disagree with, but I assumed his invitation to such a disagreement was genuine. You can (if you want) read the full response, but basically my objections were (i) that Tom confused Chaos (or randomness) with the Complex in terms of interventions, (ii) that his statements about running away from chaos (and several other statements) were simply false and (iii) more importantly that he was, as I had with Boisot, creating a new model which needed a new name while acknowledging its source. It was this latter point which caused the issue as Tom really wanted to use the name. This quote from Tom is revealing in that respect:

“On Cynefin itself, my firm opinion is that, in its own stated terms, the one who’s “moved so far away from the concept” of Cynefin is you, not me.”

I don’t know if claim this was an emotional attachment or had a commercial goal. Whatever Tom finally agreed avoid the name and on that basis I provided a detailed and polite response to the various statements he had made about Cynefin and the work of Cognitive Edge. You can find those in the final comment of the above referenced attachment.

Interestingly Tom has never responded to those comments despite several invitations, but has continued to make the same false statements about Cynefin and Cognitive Edge practice. This is different from criticism, which is always welcome. Instead its an example of the straw man fallacy in which you misrepresent your opponents position. To do that once can be a simple mistake, however to ignore the corrective response and perpetuate the error is indicative of a wider ethical problem related to the integrity of the perpetrator.

So I was disappointed by Tom’s failure to respond, a disappointment that was compounded by a series of blog postings by him in what he (revealingly) called the Beyond Cynefin Series. This got really silly at times with him looking up and using alternative welsh words, using phrases like non-Cynefin and the like. I made the  odd attempt at intervention but basically left him alone.

However in May 2010 I got involved again. Tom published a book in which he effectively claimed the Cynefin name again and also created a hotch potch of model mappings. The most absurd took the DIKW model and overlaid it on to his not-Cynefin model. I’ve replicated it as the second figure in the panel to the right and tweeted it at the time. I made it clear to Tom that if he had used any professional publisher I would have formally complained, but as he was self publishing appealed to his integrity and got a positive response. He stopped publication and changed the model to that shown as the final figure in the panel to the right. I still thought he was being a little derivative but live and let live. I made a note at some stage to make some comments on the more interesting aspects of his work such as Figure 29 shown and also another spectrum representation which showed a range from an instance of one to many.

Unfortunately Tom really can’t get away from some need to have a degree of ownership over the Cynefin framework. In Jun 2010 he referenced one use of Cynefin in systems architecture by Andrew Johnson and then revisited a lot of the old false statements from February, again showing an indifference to my prior response which evidently failed to support his personal narrative. I tweeted a comment that I couldn’t understand how anyone who had read the Kurtz/Snowden paper on Cynefin and Johnson’s paper could possibly come up with the conclusions he had and asked my twitterverse if I should both to respond or not. The suggestions ranged between “Simplistic Gibberish, not worthy of a response. Do not offer it any credence through response.” and ” Yes please respond to this weblog.tomgraves.org. I don’t understand his interpretation of Cynefin… we need clarification here.” I responded as before and Tom added ad hominem to his previous logical errors, blocked me from following on Twitter and has deleted any recent comment on his blog. He reminded me a bit of Uncle Jerry’s tantrums at this point as its not really a response you would expect from someone whose original post stated:

Dave and others may disagree with, sometimes quite vehemently – and that’s okay, because this is definitely a ‘work in progress’, and probably with gaping holes in the reasoning in places. I need that critique if this is going to work in practice.

As I say its sad, but its the nature of beast. I’ve previously referenced the story of the falcon and the pigeon from the tales of the Mullah Nasrudin and my original response was going to be along the lines of people making my complexity based falcon into a pretty linear and over structured pigeon with new age mysticism thrown in for good measure (Tom used the magic word at one point). However the tantrums of the last few days lead me to the conclusion that the pigeon has white feathers and that Captain Blackadder should be called in.

But taking it seriously ….

Ignoring the tantrums, there is some serious material to be handled in the work Tom has done. I may provide a more detailed response in the next few days, but for the moment let me highlight what I think are the main areas of concern.Screen shot 2010-06-06 at 12.07.16.png

  • Tom is using language such as chaos and complexity in a way which does not really conform with the literature. Chaos is randomness, agents operating without constraint. Principle based interventions (more commonly known as heuristics) are one of the main strategies for handling complexity as opposed to rules. Tom’s model would work better, and have higher utility if he used different language
  • In his book Tom attempts to overlay virtually every model be has come across onto his two by two matrix (OK its got a circle in the middle, but that actually means everything its not a distinct domain). This includes representing the OODA loop as movements between the domains and also the DIKW overlay. This produces a form of messy incoherence.
  • Interesting Tom, like Zachman likes structured taxonomies. He talks about the Cynefin categories (my emphasis) and pays no attention whatsoever (I suspect he lacks awareness) of the emergent nature of its construction in practice. He is imposing a structure, rather than allowing a structure to emerge.
  • He doesn’t understand messy coherence and the necessary interaction between identities within a system, those identities being people, cultures, narratives, technologies etc. We even get allocation of roles to domains (from bottom right to bottom left he assigns responsibility to Priests, Scientists, Technologists and Artists). Roles transcend domain boundaries in Cynefin they are not limited by them. Tom is confusing the system with the nature of agents.
  • In general Tom’s approach is dichotomous not dialectical, although sometimes presented as a spectrum or a set of categories. This is exemplified by the two by two labels of externalised (objective) – internalised (subjective) and also value opposed to truth in his dominant representation (shown above and to the right)
  • Tom completely fails to understand the importance of the catastrophic fold between simple and chaotic in Cynefin. Its important and relates strongly to time based concepts of catastrophe theory and issues of risk.

As well as that we have some areas of interest, triggered by Tom’s work that need development

  • Handling instances of one is a major issue in research as well as action. There is a body of literature here including March’s famous Learning from samples of one or fewer in Organizational Science Feb 1991. Instances of one are not necessarily unique, but they have to be handled as such
  • Understanding that randomness or chaos used deliberately has power. Cynefin and Cognitive Edge have always had methods in this domain that are not simply about rapid exit (Tom’s major misrepresentation) but are linked to innovation and mass consultation. If we use language properly here then a shift into chaos is a temporary and contextually located total relaxation of all constraints.
  • While the dichotomy between value and truth (or magic and fact) in Tom’s world is dangerous. However it is the case that the development of value systems can be a key factor in managing risk in a complex environment and one way to prevent and unplanned entry into chaos. This links to work on communication based on metaphor which is more powerful in conveying intent and what I will call value laden intent that more formal methods.
  • Its probably time to start publishing some of my (and others) thinking of different aspects of complex systems where humans are involved. Cynefin splits ordered systems into two, simple and complicated based on human perception. I have been playing with a similar structuring of the Complex domain that would relate to the degree of coherence and vector control possible in human intervention.

In practice I think I will develop those ideas in future posts rather than further respond to a Tom who’s future engagement is really in his hands. For his own sake I hope he learns to read considered responses to his comments and then either to provides further explanation or argument, or just to admit a failure to read or understood something properly. In the meantime, and for the avoidance of doubt his model is no longer even loosely connected to Cynefin other than a superficial use of language, and his own obsessions. If he can shift from being a confusionist to something more coherent we may get somewhere.