« Style | Main | Civilisation »

Confusing symptoms with cause

An interesting set of presentations down here in Tampa, Florida over the weekend. More reflections over the next week, but one of the things that came up in questions reminded me of the transformatory nature of complexity thinking. Someone asked me how you created the trusted environment prior to engagement. My response (including some follow ups) was two fold:

  • that trust is an emergent property of the process of engagement not a precondition. You can't design for trust, and attempting an overt discussion of the issue in most organisational environments just produces linguistic conformance; picking up on management and facilitator language and replaying it. Accordingly rather than talk about the idealised or desired qualities of the conversation, you need to create environments and interactions where it is more likely to emerge.
  • that trust was not the same thing as being open to others ideas, it could include the ritualised attack or criticism of ideas or the assumption of a contrary position for the purpose of increasing the range of perspectives taken into account. I Use ritualised dissent to great effect for example. Groups work in parallel on the same issue and then a spokesperson presents the ideas of one group to another, received in silence. They then turn their chair and their ideas are subject to severe criticism to which they are not permitted a response. Multiple rounds of this increases the resilience of any plan or analysis and avoids the fragility of premature or unthought consensus.


Now this is not conventional wisdom in idealised forms of facilitation. However it uses complexity principles: increasing interaction to break down existing assumptions, preventing premature convergence, increasing the diversity of agents. I could go on but then this would become a 101 complexity course.

The trust question is a classic confusion of symptoms with cause, just as creativity is a symptom of innovation not its cause, so trust is the symptom of interaction over time. If that interaction is not testing, then the trust is fragile. If the trust is simply the result of few contextual exercises (throwing yourself backwards off a brick wall is the classic) then it is temporary. Focus on the process, rather than trying to preset emergent outcomes and you get a more sustainable solution,

Comments (9)

Brian Sherwood Jones:

I'm not so sure; At first sight, I thought that Bob Sutton's 'No Asshole Rule' was simplistic since it didn't really change the system that created and nurtured assholes. Now, I am coming round to it, and maybe a good bash at that with support from the top would be a useful precursor to almost any intervention that depends on honest input from those lower down the pyramid.

Dave Snowden [TypeKey Profile Page]:

Not sure of the point Brian. Support from the top never does any harm, but its one input to the evolution of a complex system. You still can't manufacture trust, it evolves

WalterRSmith:

Dave:

Good comments on an important topic...part of the problem I've had when discussing trust is having a clear working definition of the word.

When discussing decision making, I've found it helpful to emphasize two aspects:

(a) Within the target context, the degree to which the other person/group makes sense of the context the "same" way you do, and

(b) Within the target context, the degree to which the other person/group has the will to act consistently with the sense that is made of the context.

[Note: "same" = roughly similar in regard to key decision making factors]

The first item goes back to shared narratives, values, goals, etc. Both items point to "sources of trust", which as you point out, accrete with time/experience.

I'd be interested in other aspects of trust you or your readers would emphasize. BTW, I'm mostly focused on trust in a collaborative context....not an adversarial context...not that the two are necessarily disjoint.

Brian Sherwood Jones:

No, manufacturing trust is like faking sincerity. However, there are specific things one can do to reduce untrustworthy behaviour. Bob Sutton quote from someone at Google "I am really not a very nice person, but I have to act nice here, or I can't get anything done."
So, yes, one input to evolution, but a specific one, and it is my impression that Sutton has quite a bit of evidence that it works. It was my speculation that doing something on that early on might help other activities.

Dave Snowden [TypeKey Profile Page]:

I must admit that the Google quote disturbs me, and reinforces the issue of linguistic conformance I mentioned. I think Bob Sutton over emphasises individual over system change, and like all the popular authors it is easy to assemble evidence with the benefit of hindsight: it is rarely a statistical sample of the available data! Always a good read however.

Walter - I agree that shared narrative context is key, but I think ritualised conflict is as (if not more) important than collaborative discussion which carries the danger of said linguistic conformance.

Nat W [TypeKey Profile Page]:

Is defining trust much like defining pornagraphy? As a US Judge once said, "I can't give you a definition, but I know it when I see it".

Nat W [TypeKey Profile Page]:

Is defining trust much like defining pornagraphy? As a US Judge once said, "I can't give you a definition, but I know it when I see it".

geoff elliott:

Interesting comment but didn't Deming cover this ground some 20 odd yaers ago, ie common and special cause. Surely we should be drawing upon the work of Deming and not trying to reinvent something. Not also the work on TRIZ 99.9999% of ideas are not new just a restatement of something elase - so what is innovation?

Thoughts

Geoff Elliott

Dave Snowden [TypeKey Profile Page]:

TRIZ is interesting but makes assumptions about semantic analysis and some of its applications (white space patenting) seem to me to be anti-innovation if anything. Overall I think the field has moved on form this (and certainty from Derning) although there is still learning

Post a comment




Remember Me?

(you may use HTML tags for style)