Verna Allee responded to my request for help with some interesting reflections on high abstraction languages. She specifically raised examples from Meg Wheatley and the Dillema Theory of Charles Hampden-Turner. My response covers issues on archetypes and high abstraction language,and some of my concerns at therapy based approaches to organisational studies and the way in which people with systems-thinking approaches such as Wheatley and Hampden-Turner use the language of complexity to support an older paradigm: trying to put new wine into old wineskins (Matthew Chapter 9 verse 17) – well it is Sunday again)
Verna and I agreed on several issues, but on one key point strongly. In respect of any sense-making method she says: the key qualifier is whether it is something that unskilled people can learn, comprehend and even replicate with just few hours of learning
VERNA’s ORIGINAL POST =============================================
I am intrigued with your statement: ” we established that describing a situation at a higher level of abstraction than normal improves not only weak signal detection (seeing things that might otherwise not be seen until too late) but also resilience to manage change. In a sense that finding was counter intuitive, but supported by other research that shows more data and lower levels of detail do not improve sense-making.”
Years ago Meg Wheatley made a remark that what we really need in our organizations is the ability to see patterns. I believe you are exactly right that with complex adaptive systems there is great value in describing a situation or system at a high level of abstraction as a way to see patterns. I think of these exercises as basically holding up a mirror so the system can see itself. Every living system want to move toward health and wholeness and these mirrors allow the intelligence of the system to know itself and thereby begin the healing process. Sense making at this level has been completely underappreciated in business management practices, yet it is an essential capacity for adaptation and response.
It would be a great contribution to define a core set of descriptive processes that support sense making at this level. Your approach to storytelling and describing archetypes is clearly one of those approaches. The value network modeling is another. Each of these suggests different lenses to understand the behavior of the actors – in the storytelling thinking about psychological or cultural archetypes evokes that shift of thinking. In value network analysis thinking about roles and deliverables evokes a different kind of pattern senseing.
There are others of course – we could probably brainstorm quite a list in a very short time. However, for me the key qualifier is whether it is something that unskilled people can learn, comprehend and even replicate with just few hours of learning. Perhaps the simplest is the After Action Review. The sense making there emerges in the tension between what was supposed to happen and what actually happened. It is simple, elegant and fast to learn. The basics of VNA can be mastered in a few hours because you have three elements of nodes, directional arrows and labels that represent deliverables. Techniques for conscious conversation with world cafe approaches can also be mastered very quickly.
So my suggestion is to focus on the most elegant, powerful and easily learned pattern sensing tools and simply validate their usefulness.
As far as problems, the liveliest conversations seem to be triggered around issues of trust and governance – the tension between structure and freedom. Framing a good creative core set of questions around those issues would also be very helpful. I like the way Charles Hampden Turner frames such questions as the two horns of a dilemma where the real question is how to live with the dilemma rather than solve it.
MY RESPONSE ===================================================
An interesting response Verna. I would like to agree in part and then provide some qualified disagreements. In respect of the latter I want to exaggerate differences a bit to make a point as I think you have brought into play two people who exemplify a systems thinking rather than a complex systems perspective: Wheatley & Hampden-Turner. None of this is really fully justified by your post by the way, but you triggered several chains of thought.
By the way, something about the Welsh (or at least my tribe). We operate a simple rule
If we like you we argue with you
If we are unsure about you we are polite
If we are very polite you will shortly die or we will die in the attempt
Before going there I want to take a brief look at the statement (which which I agree) that archetypes and networks representations provide a high abstraction language. Some of the text that follows assumes that readers know the archetype method which Verna refers to, If you don’t and are interested in knowing more then the full method can be downloaded from our web site under the open source license. Examples of its use, with some updates on method were published in E:CO in my Frontiers piece.
The high abstraction approach to sense-making is an interesting one. Our work on archetypes as emergent properties of complex interactions was one early experiment in sense-making which persists in both value and active use. It has been compromised in execution from time to time, generally for the best of all possible motives; the road to hell is paved with good intentions. One such compromise involves getting people to talk explicitly about the archetypes as a group with direct facilitation, rather than allowing them to emerge from natural processes. A lot of the consultancies that focus on creativity tend to that approach. I think because they are linked to charismatic facilitation and visionary modes of conversation. The problem is such an approach leads to stereotypes not archetypes. Another compromise is to give people lists of pre-determined attributes for clustering. It speeds up the process, but inevitably reduces the richness and introduces bias. Both of these errors place too much structure from an external agent into an emergent process and thus contaminate the field of study. When we did the original experiments on this we worked in environments where I could facilitate in English, but all the conversations would take place in Danish or Mandarin or Glaswegian (the last partly but not completely in jest) so that I could not understand the content or output. This avoided facilitator infuence or bias.
I think the reasons for both of these deviance types is that people find it difficult to shrug off five things:
(i) the Jungian & Campbellian approach that archetypes represent universals
(ii) the myth of consultant as therapist
(iii) metaphors of wholeness and health that permeate a lot of systems thinking
(iv) the general pointless concern of post modernism with ideology and deconstruction
(v) an unwillingness to have workshop participants experience uncertainty.
The idea of archetype creation is to allow the process of emergence to happen without control so that the results reflect the perspective of those being studied. Interestingly we have found to-date, that indexing systems for narrative databases that are derived from archetypes seem to produce more interesting and intelligible patterns in reporting than do those where the index sets are analytically constructed by experts. Now we are working on that, but its an interesting early piece of anecdotal evidence. The fact that the abstraction comes from the people themselves gives its more power. A good example of archetypes as steriotypes by the way can be found in Angeles Arrien’s article Four Fold Ways to Wisdom. I operate a general principle in KM by the way: If someone mentions the wisdom word, especially if they conjoin it with management treat everything they say with the deepest suspicion.
I would agree that network representations also provide a form of high abstraction representation which can be used for sense-making. I think this is especially true where network analysis tools are used to represent and simulate events that have taken place without qualitative input. I think less so where the network is based on questionnaires about relationships or qualitative not emergent assessments of value.
In the experiments I ran with Gary Klein in Singapore took place in a battle field command and control environment. We showed that we could improve two things (i) weak signal detection and (ii) resilience and speed of adaption when conditions changed. The method was to get people to address the situational assessment in terms of ATTRACTORS and BARRIERS rather than by describing what was going on: to use a high abstraction language. In addition interventions were planned around shifts or changes to the attractors and barriers, rather than looking at objectives. I will do some work next year on this in respect of effects based operations EBO in which we will look at something called short-cycle EBO. This uses that established method of planning but focuses it on attractors and barriers not on specific battle field objectives. In a weeks time I am working in an experimental environment, again in Singapore but this time with non-military environments to further develop methods around high abstraction languages. This time with Alicia Juarrero whose book Dynamics in Action is one of the best I know on complexity. This is a part of a wider set of interventions planned around issues of network government in respect of policy making and safe-fail, not fail-safe intervention strategies. This is a fruitful area of study and we hope to release some methods into the open source environment towards the end of the year so more people can experiment with them.
Now to Wheatley & Hampden-Turner.
Now I agree with Meg Wheatley that we need to develop the ability to see patterns, but I am not so sure about phrases such as Every living system want to move toward health and wholeness and these mirrors allow the intelligence of the system to know itself and thereby begin the healing process. Now I want to make it very clear that I am not against health and wholeness! However I am worried here by the assumption that a complex system has a naturally positive direction and the assumption that adopting “living system” theory means that we would move in that direction. I think that is dangerous. A complex system has no purpose or direction is simply “is”. If we start to understand that it is complex, that there is no pre-given future and that we cannot take an idealistic approach of defining or advocating a future state then we have a damn sight better chance of getting to a better future.
I think there is some hang over here from the out of date concept mental models. If we look at Senge (and he and Wheatley are linked) he argues that y”u can achieve self-awareness and purposeful of mental models. Others such as the Appreciative Inquiry types, while having some useful methods take them to extreme. They do this by assuming that positive stories arise from positive mental models. They then assume getting people to tell positive stories will give them positive mental models. A logical non-sequitur which is staggering in its stupidity, but all to common in management text books: these people are successful and they do X, so do X and you will be successful. This idea of purpose and ability of the master to direct evolution in a favorable direction (not stated but implied) is I think dangerous. You can see a form of religion emerging here which takes on a Western-USA perspective on Buddhism (very atomistic and leadership orientated) combines it with Gaia theory to create a heady, dangerous and irresponsible mix. It seems to focus on personal change (mastery) not on system change for example and in some of the manifestations I have observed carries with it many of the objects of religious observance.
A lot of methods in this space also seem to draw on the practice of therapy. One cynical reason for this is that consultants like methods based on therapy as they are then the therapist and are in a position of power. However I think the more dangerous point is that with the metaphor of healing, comes the idea that there is some natural state to which things should be restored. Some form of the ideal to which we have to return and which is represented in natural systems. This has a long history in human thinking and generally leads to withdrawal not active involvement in change.
My response to this is simple – we can return to a natural system if we give up language and stories (the two basic components of our intelligence) and act as (literally) dumb beasts. That will restore harmony and the ability to use natural systems models. Of course then no one would be around to study it, but Gaia might recover! Its a lot safer to assume that the system “is” (more Tao than Zen) and work from there without the idealistic presumption or the quasi religious language. I think that two things will work here: (i) high abstraction language and (ii) switching the strategy paradigm to multiple safe-fail not fail-safe interventions. This allows you to respond to the good with reinforcement and to the bad with disruption. Seeking the ideal has a long history, it produces many saints but few paradigm changes. An aside: one can agree with Kuhn on paradigm shifts without agreeing with his view of science as a whole.
Hampden-Turner frames questions as dilemmas and places people in a condition where they are asked to live with or resolve the dilemma. In practice the method leads to solving rather than living with. It is the nature of a dilemma that we have to make choices, Sophie’s Choice being to my mind the worst I can imagine, and the language leads to that. Now because systems-thinking is about causal systems (all be it very complicated ones) one can see how this position arises. However in this area I prefer to focus on PARADOX over DILEMMA. The nature of a paradox is that you cannot make a choice – but that you are forced to think differently about the world.. I am not sure that talking about trust and governance is as effective as making small changes to see if trust and governance strucutres change. A great example here is the Grameen Bank.
Now I fully agree with you on this – brilliantly summarised if I may say so – the key qualifier is whether it is something that unskilled people can learn, comprehend and evenreplicate with just few hours of learning. This is vital. Some points in relationship to this
(i) we are finding that tools which pass this test in the field with people under conditions of uncertainty often fail when reviewed by middle management or in case based training where there is little of no uncertainty.
(ii) Narrative based after-action-review I think passes the test. However the minute people introduce facilitation, structure and analytical frameworks (which they do) then it breaks down
I would therefore modify your statement (less elegantly) as follows:
the key qualifier is whether it is something that unskilled people can learn, comprehend and even replicate with just few hours of learning, or ideally none if the language or model are intuitive; such methods should not require facilitation or expert interpretation.
Thanks for a great post and apologies for the length of the response but it was a simulating start to the day here in Singapore – now I must get back to the book. However I will post this and your original on the blog before doing that.
Cognitive Edge Ltd. & Cognitive Edge Pte. trading as The Cynefin Company and The Cynefin Centre.
© COPYRIGHT 2022.
My thanks to Paolina via the ACT KM Forum for bringing my attention to this ...
Today’s Observer has a lead article with a very scary message. I quote:.THE TRUE scale ...