Dave Snowden

Of experts and expertise

RSS Feed


A long time ago when I first created the Cynefin framework in its five domain form my focus was on knowledge management.  The first fully worked article looked at communities and deployment of knowledge and in those versions I talked about complicated (then called knowable) as the domain of experts.   The context here was the then tendency to assume that all you needed to do to create a KM programme was create a taxonomy and set up a CoP programme.  CoPs were then held to be the best way to host and develop expert knowledge, so calling it the domain of experts made sense.

Most people understand that to say that one domain is that of experts does not mean that expertise is not required in the others.   The point is, and has always been, that in the complicated domain you can trust expert judgement.  In the complex domain you need conflicting expertise as increasing scanning is needed and the danger of expert entrainment to known or knowable contexts means they will not spot weak signals and/or that the context has shifted.   We have the odd anal individual out there who argues that if you use an expert it means it must be complicated.  Aside from a failure to appreciate context, such misguided souls really require a quick lesson in remedial logic; to say that condition A requires the application of B does not mean that if B is applied condition A must aways pertain.  It’s not just poor logic, but also represents an unhealthy need of some consultants to everything in neat little boxes with little ambiguity; a need for the simplistic that is worrying in people who purport the competence to advise others in ambitious situations.

Of course expert bias is an issue where you are dealing with novelty or changed context as this quote from The invisible gorilla strikes again (
Drew, Vo & Wolfe 
 Psychol Sci. Sep 2013; 24(9): 1848–1853).

What about expert searchers who have spent years honing their ability to detect small abnormalities in specific types of image? We asked 24 radiologists to perform a familiar lung nodule detection task. A gorilla, 48 times larger than the average nodule, was inserted in the last case. 83% of radiologists did not see the gorilla. Eye-tracking revealed that the majority of the those who missed the gorilla looked directly at the location of the gorilla. Even expert searchers, operating in their domain of expertise, are vulnerable to inattentional blindness.

Nothing there says that you should not trust a radiologist if there is a risk of lung cancer, the point is that the patterns of expertise require focus and that focus itself can become a form of blindness.   At an organisational level this is the point made by Clayton Christensen who put forward the idea that success blinds competent managers by virtue if their competence.

So in the complicated domain we reduce overall energy needs by placing reliance on expert judgement, in the complex domain we deliberately introduce variety of expertise and even naiveté to ensure wider scanning with a greater focus on anomalies.  In the obvious domain expertise is embedded in process or embodied in training.   The chaos the expertise lies in creating the condition or applying constraints.