REPLY Policy paradox and SD (SD6689)

[SDMAIL Jack Harich]

Posted by  Jack Harich <register at thwink.org>


One way to handle the implementation aspect of the "solution" the model 
suggests is to not look at it that way. There is another viewpoint.

This is to first decompose the one big problem into several little 
problems, each of which is much easier to solve. The decomposition I've 
found most successful and reusable is these three subproblems:

A. How to overcome change resistance to adopting the proper practices of 
the "solution." Think of the system with the problem as suffering from 
organizational change resistance. When change resistance is overcome the 
system will want to solve the problem just as strongly as it was not 
wanting to solve it before.

B. How to properly couple two subsystems to another. Proper coupling 
means the feedback loops between the two systems cause the systems to 
work together in harmony.

C. How to avoid model drift. The "model" here is the model of the solution.

Call A, B, and C the change resistance, proper coupling, and model drift 
subproblems.

In general you first solve A, then B, then C. However, your analysis 
iterations will probably initially center on B, as you become familiar 
with the problem.

Regarding C (model drift), as the total system with the problem evolves, 
so too must the model, or the solution will become obsolete. This is 
also called a self-managing solution. The term model drift comes from 
Thomas Kuhn's cycle of scientific revolutions. The cycle steps are 
normal science, model drift, model crisis, model revolution, and 
paradigm change. I added the model drift stage, which was only implied 
in his book.

Regarding B (proper coupling), suppose the problem was the diabetes 
epidemic. The two systems (there are sometimes more) could be defined as 
citizens and the educational system. The educational system needs to 
constantly incorporate new methods based on what youth needs to learn in 
its formative years, such as the right diet and exercise necessary to 
avoid a susceptibility to diabetes. The educational system also needs 
feedback on how well its working. When these things are happening the 
two systems are properly coupled. Another example is the environmental 
sustainability problem. There the human and environmental systems are 
improperly coupled due to the externalized cost phenomenon, though this 
is an enormous simplification.

This takes us to A (change resistance). Implementation may encounter 
organizational change resistance, a barrier so common it's a 
mini-industry in itself. So in your analysis you model change resistance 
explicitly, usually in a separate model from B and C.

Generally the more difficult the problem, the higher the change 
resistance. For incredibly difficult social problems, like 
sustainability, systemic poverty and war, change resistance is the crux 
of the problem. That is, we know what to do to solve the problem (the 
proper coupling practices) but we cannot get the system to adopt those 
proper practices.

Thus if you have not solved A, then no matter how good your "solution" 
to B is, you have not solved the total problem. Most people consider 
only B as the problem to solve, which, if it's a difficult social 
problem, is the perfect setup for a trap. When they try to implement 
their solution to B, the one they have worked so hard on for so long, 
they discover the system rejects it with surprising force and ingenuity.

An educational example of A, B, and C may be seen in Forrester's urban 
decay model. The model showed how certain improper coupling practices 
caused urban areas to be improperly coupled to the world around them. 
This caused urban decay. The model also showed how, if new counter 
intuitive policies were tried, proper coupling would occur.

But when Forrester started publishing, he encountered severe change 
resistance. For example:

"The conclusions of our work were not easily accepted. I recall one full 
professor of social science in our fine institution at MIT coming to me 
and saying, ‘I don't care whether you're right or wrong, the results are 
unacceptable.’ So much for academic objectivity! Others, probably 
believing the same thing, put it more cautiously as, ‘It doesn't make 
any difference whether you're right or wrong, urban officials and the 
residents of the inner city will never accept those ideas.’ It turned 
out that those were the two groups we could count on for support if they 
became sufficiently involved to understand. That is a very big ‘if’—if 
they came close enough to understand. Three to five hours were required 
to come to an understanding of what urban dynamics was about." - From 
http://sysdyn.clexchange.org/sdep/papers/D-4165-1.pdf

In the urban decay crisis problem, change resistance was so small that 
all it took to overcome it, one group at a time, was "three to five 
hours" of exposure to the model. Rather than explain the whole model, 
all Forrester explained was why present solutions had no effect or made 
the problem worse, and how alternative solutions that had not been tried 
could make it better and why.

But Forrester and others never addressed C, the model drift subproblem. 
The urban decay crisis was merely resolved down to a problem. It was 
never entirely solved. We still have slums all over the world. We even 
have occasional riots, such as those recently seen in France. But we do 
not have anything like the US urban decay crisis of the 1960, when 
scores of people were slain in riots, such as 34 in the Watts riot alone.

As another example, a friend is a business consultant. He uses a 
repeatable process for assessment, strategy development, etc. He feels 
that most business problems fall into three categories: change 
management, project management, and communication. From the perspective 
of A, B, and C, his change management category is A and project 
management is B.

Now you may wonder, what about C? Well I asked him, what about strategy? 
Isn't that missing from your categories? He replied "That's a whole 
'nother story." But I feel that if he had a fourth category called 
strategy management, it would be C. That is, for a system to avoid model 
drift, it must continually fine tune its strategies, the ones that lead 
to continual updating of the solutions to A and B.

Hope this helps,

Jack Harich
Systems Engineer
Posted by  Jack Harich <register at thwink.org>
posting date  Wed, 12 Dec 2007 11:21:40 -0500