Dave Snowden

Description not evaluation

RSS Feed

One of the things I am enjoying at the moment is the way a lot of things are coming together conceptually.   It happens like this when you develop or repurpose knowledge from different sources.  Individual practices and ideas make sense in their own right.  Then you read more, practice more, and the various origins and instantiations (sorry can’t think of a better way of saying that) then synthesis into something more coherent.  I’ve found it is the nature of trans-disciplinary work and its important not to force the synthesis.   That, in complexity terms would be the danger of premature convergence.  Holding things open, allowing them to break down into more finely grained objects, then seeing them recombine and co-evolve always reminded my of sitting in the Hilton in Rome on the path to a Welsh Grand Slam in 2005, seeing the murmuration patterns of starlings over the dome of St Peters.

I found that happening over the past two weeks linking key issues in scaling complex adaptive systems with the separation of description from evaluation, or description as evaluation; something I have increasingly realised is a key part of SenseMaker® as well as methods like Ritual Dissent.   The whole point of SenseMaker® is to transfer the power of interpretation to the subject of research this making them (sic) a subject not an object.  When we do that we find that comparisons between their self-signification and signification of the same material by experts reveals an interesting difference.   Left to themselves people, taken as a cluster or series of clusters, describe their situation while experts interpret it with a high level of instrumentalism.   That is important in terms of cultural bias and power but it is also vital in terms of discovering novelty.

The longer we can carry on describing things without judgement or interpretation, the more we are open to possibilities that we might not have seen at the time.  It is the theory behind the cliche of if only we knew what we knew quoted by many a knowledge manager.  They tend to try and codify which is always evaluative and thus loose material.  True description allows you to find solutions that are screened out by evaluation or interpretation.

Humans by their nature tend to instant judgement then the evidence thereafter confirms that judgement or intuitive bias (which is not the same thing).   So the workshop techniques I was teaching last week move people around between groups not allowing a pattern to stabilise.  In contrast with Open Space you cannot choose what interests you as you might need to confront something that would normally convert the law of two feet into the practice of a clean pair of heels.   As each part of a process takes place with different actors we hold open our interpretation for as long as possible, and even when the focus is on actions they are finely grained, working in parallel and subject to multiple rounds of ritual dissent. Even the actions become descriptive when practiced in this manner.

Similarly in SenseMaker® multiple agents responding in real time, coupled with the ability to ask questions of that network (I will blog on that in a few weeks time) before an issue is substantive, allows early intervention and multiple re-descriptions from different perspectives.  Multiple human agency augmented but not replaced by technology is key here.  It also means the three basic heuristics of complexity are in operation in their full glory.   Finely grained objects, distributed cognition and disintermediation.

The more I work in this field the more the whole thing distills down to simple principles applied in different contexts to create novel but coherent approaches.  The new simplicity that takes a conceptual leap and a practical experience or two to achieve.   Focusing on description not evaluation or interpretation is key.   This allows changing descriptions to become a form of evaluation (monitoring AS evaluation) and also means that judgement can be the emergent property of multiple interactions both direct and virtual with multiple actors over time.   The human interpretation at the start of the process (self-signification) and at the end (what do we do now) is key, automation is about augmentation not about producing automata.