REPLY Limits to Growth Plan B (SD7031)

[SDMAIL Jack Harich]

Posted by  Jack Harich <register at thwink.org>

Jack Harich innocently asked:

So, how would you go about solving the [environmental sustainability] 
problem?

Assume you had 105 million dollars per year, which is the United Nations 
Environmental Programme's annual budget.

Mike Fletcher courageously replied:
> Jack:
>
> That is a legitimate question, although a pretty hefty one to address 
> fully in one forum post. I  don't claim to have all the answers. I 
> freely admit my "model" is far from complete!  Perhaps I can 
> contribute a bit by asking a few key questions.
>   

Mike - There are lots of good ideas in your post. I don't have all the 
answers either. Let's treat all this as an exploratory educational 
exercise. Here goes:

> Firstly, 100 million dollars seems to be a rather severe case of 
> underfunding if your goal is to change fundamental ways of thinking on 
> a global scale.  

How would you define the goal? Would it be changing the way people 
think, or how the system behaves?

> Without being in the least facetious, a viral marketing campaign is 
> about the only hope with funding like that.
>   

Where is the analysis that supports this fundamental conclusion?

> If we are looking for the most fruitful way to spend our time and 
> money we need to start somewhat cautiously.  As I warned in my 
> original post,  before we jump into the solution phase we have to do a 
> better job of problem definition - and frankly I think that "the 
> problem" has largely been framed in a solution-driven way, and most of 
> the solutions have been focused on the symptoms and not the illness. 
> But we don't have the luxury of admiring the problem for another 30 
> years either.
>   

Yes, a good problem definition is crucial. I have a sample one in my 
execution of the System Improvement Process. That process says that 
difficult complex social system problems are best defined starting with 
this standard format: Move system A under constraints B from present 
state C to goal state D by deadline E with confidence level F. Without 
going through the details, the result can be summarized as:

The global environmental sustainability problem will be solved when all 
critical environmental properties are being held in their safe zones 
indefinitely or are moving there within a predictably safe time span, 
with a confidence level of 99.9999%./

/The high confidence level arises from wanting the solution to hold for 
a minimum of 10,000 years. I think we can assume the solution will be 
replaced by another one by then. You may recognize the six nines from 
Six Sigma.

I like your observation that "most of the solutions have been focused on 
the symptoms and not the illness." That's exactly what the standard 
problem definition format was designed to avoid. Proper focus is so 
important.


> So, what are the solutions to the limits to growth problem? Hopefully 
> there aren't too many people left who would dispute that oil 
> depletion, global warming,  water crises,  population, pollution etc. 
> are all linked issues and really not separate problems. 

Yes!


> Hopefully the size of the group that still thinks these issues can be 
> solved independently is also shrinking.
>   

But yet the world IS trying to solve them independently, apparently 
starting with climate change as the number one priority, and various 
types of pollution (generally local) as number two. Hmmm....

Nice thoughts.

> I don't think that anyone would dispute that these linked issues 
> definitely fits into the category of problems described as "wicked 
> problems."  Before we start proposing solutions we should probably ask 
> one of Polya's questions, "Under what conditions can I solve the 
> problem?"  What little we know about such problems in general is that 
> "solving forwards" (classic waterfall approaches) show poor results.
>   

The consensus, at least in the quality and software engineering 
industries, is that failure of waterfall approaches is mainly due to 
insufficient iteration and attempting to do too much too early. See:
http://en.wikipedia.org/wiki/Waterfall_model
 From the above link:

"David Parnas, in 'A Rational Design Process: How and Why to Fake It', 
writes:

“Many of the [system's] details only become known to us as we progress 
in the [system's] implementation. Some of the things that we learn 
invalidate our design and we must backtrack.”


> Essentially,  "solar power solutions," "hydrogen solutions" etc. are 
> clear example of "solving forwards" approaches.  It is not clear (to 
> me at least) that they even address the real problem, let alone are 
> effective solutions. 

I'd go so far as to say they are not even waterfall process results, but 
something else. They are intuitive quick fixes (based on no process at 
all or Classic Activism) to tiny parts of the problem. They have a very 
low success rate so far. Some even make the problem worse, like the 
recent infatuation with biofuels in the US and elsewhere.


> While models could be used to test such approaches to some extent, 
> there is a real risk that such models would fall prey to what amounts 
> to the logical fallacy of affirming the consequent.  The stakes are 
> too high; we should avoid such plans to solve the problem as a matter 
> of course, simply because the they are unlikely to lead to optimal 
> results.  Grabbing the first available apparent technology solution 
> (or hodgepodge of available technology solutions) and running with 
> them is an easy but perilous course, and certainly not a strategy I 
> would like to stake the future of the planet on.  

Yes. I see what you're saying: In general, popular solutions have 
centered on quick, intuitively derived conclusions about what will work. 
Proof that this will not work lies in the history of how well they have 
worked since the problem was first identified in 1972 by the LTG project.


> Clearly there are many people who have moved beyond that kind of 
> thinking, but many remain who think we can simply "technology" our way 
> out of the problem.  The technology as deus ex machina assumption is a 
> huge issue in how the world is currently addressing the problem. This 
> is covered in Chapter 6 of Limits to Growth, so I won't dwell on that 
> aspect of the problem overly long, except to say that solutions driven 
> by that set of assumptions about the utility of technology alone to 
> solve the problem could lead to "false spring" outcomes or worse.
>
> Due to the perils of the above I think that  "solving backwards" is 
> really the only realistic plan, and in that modeling could help quite 
> a bit.  Let us propose a model of the future that could work and work 
> backwards. Lets ask a series of questions, some of which  that can be 
> addressed by Modeling and Simulation (M&S).  Clearly M&S is really the 
> only way such complex proposals can be evaluated.  We really need to 
> "clear the decks"  in regards to initial assumptions.  Nothing would 
> be off the table, initially at least.  As someone once said, most 
> ideas progress from "that's crazy!" to "everyone knows that!" in 
> mysterious ways.  It's not much of a stretch to say that because of 
> the tremendous depth of assumptions surrounding this issue, that the 
> most "obvious" solutions to the problem are probably the least likely 
> to survive objective scrutiny.
>   

Yes to a gaggle of assertions. You're covering a lot of ground! :-)

> For example, my current working hypothesis about what definitely won't 
> work and what might work looks something like this: What can't work: 
> There is NO sustainable high-energy consumption future. Such societies 
> are inherently unstable. Simply stated, we move too fast for the 
> planet to adapt to us; too much commotion, we need to slow down.  What 
> might work: A low-energy, high-tech society that survives in 
> relatively small largely self-contained economies where people and 
> goods move minimally (too much energy cost otherwise) but information 
> moves freely.    

Working backward from future scenarios is one way to drive an analysis. 
But how do you productively generate all the scenarios that need to be 
examined? This is a solution space search based approach. It starts with 
listing all (?) possible solutions, and then examining the ones "worth" 
examining. This is the classic brainstorming approach.

The weakness of this approach is one can easily generate tens of 
thousands of possible scenarios. This is identical to the way a doctor, 
when confronted with a patient who has a fever, could easily list 
thousands of different treatments.

Now then, what criteria would the doctor use to whittle that long list 
down to a manageable one? Why of course - the doctor would diagnose the 
cause of the illness first. Even if it was not a definitive diagnosis, 
the list would shrink to less than a hundred options, probably to less 
than ten. Now the doctor has a solvable problem. Further diagnostic work 
and comparison of the remaining treatment options would shrink them to 1 
or several main ones, with an order in which to try them.


> The questions - in approximate order of the asking are: Is there any 
> structure and associated decision rules which can lead to eventual 
> acceptable equilibrium under any conceivable conditions?
>   

Again, "is there any" generates a huge list of solutions in the solution 
space, if I understand you correctly.

> Sub-question, can we even obtain a measure of equilibrium?
>   

The problem definition did this. Notice how I'm using reusable standard 
terms, like Problem Definition. This greatly reduces meandering around, 
into less productive efforts.

> Is there any structure and associated decision rules which can lead to 
> eventual acceptable equilibrium given reasonable assumptions?
>   

This is another "is there any" question.

> Is are multiple structures (perhaps quite different) which lead to 
> such equilibrium? What do these different structures have in common? 
> How do they differ? How robust are they under a variety of assumptions?
>   

This "is there any, how to they behave" approach has the limitations 
discussed above.
> We need to spend quite a bit of time evaluating alternatives 
> objectively. Despite the time spent, we cannot risk focusing too 
> quickly on the implementing the "obvious" solutions. To a large degree 
> I think that has been the general focus of most problems solving 
> attempts in this arena, but that is generally a rather poor approach 
> to getting optimal solutions to complex problems.
>   

Yes

> We must also consider what decision-rules and underlying 
> ways-of-thinking that must be adopted in order for these "models" to 
> work. (That is, not fail due to policy resistance or inability to 
> implement)  Under what conditions could the needed changes in thinking 
> happen? Are they realistic? What changes in physical structure are 
> required? Which would require the most restructuring and which would 
> require the least?
>   

This appears to be an elaboration of the "is there any" approach.
> Plan B and other proposals very generally fit into these type 
> categories of possible plans. They are however for the most part 
> untested hypotheses; that is, could the structure they propose even 
> lead to the results suggested? M&S could attempt to answer those type 
> questions.  Problems like this simply have to be tested through 
> simulations.  Testing of such models would at least allow for 
> discussions about solutions to take place in a way where solutions can 
> be challenged and tested and thus improved.
>
> Generally speaking any solution will flow from and be created by 
> changing the fundamental goals of society, and as I stated the biggest 
> leverage point by far is changing ways of thinking. 

See my earlier objection to this key assumption. But don't worry. This 
is an educational exploration.

> Perhaps the biggest single win in this regard is to change the 
> time-horizons of our thinking dramatically. This is happening but 
> perhaps too slowly. Societies are starting to think along the lines 
> that their time horizons need to add a zero - that is go from 3-5 
> years to 30-50 years.  If we think about the relevant time constants 
> involved even that appears too conservative. Time Horizons for 
> planning how humans interact with the earth probably need to add at 
> least two zeros and perhaps three.  The time-constants of the earth 
> and the time-constants how human plan their interaction with the earth 
> are off by at least two orders of magnitude. This because the "hidden 
> hand," while it works well for short-term optimization, it is 
> essentially blind to long time-horizons. 

> The "hidden hand" is also largely blind to any concept of a shared 
> responsibility for the commons.  There is a possibly apocryphal story 
> of a Minnesota Native American Tribe who identified their land-use 
> time horizon as the life-span of the major pine tree on their land - 
> 300 years, if I recall correctly.  A great story - which might even be 
> true!   In my mind, people who adopt that way of thinking are closer 
> to success than those who spend their days deciding whether to fund a 
> billion to Fusion Research or a billion to Hydrogen Fuel Cell 
> research.  Again, as I stated, it is extremely tempting, but very 
> perilous course to attempt to "technology" our way out of the 
> problem.  Recall that the problem was largely caused by the 
> unconsidered use of technology!  The rough outline of a framework 
> which has better potential answers is pretty obvious - consider more 
> and technology less.
>   

Yes. Shared responsibility, long term thinking and avoidance of 
technological optimism are mandatory. These are good concepts to keep in 
mind as we go about solving the problem.
> At any rate, there is partial attempt to answer your question. I'll 
> end here because I think I've tested the reading patience of the 
> list's members sufficiently for one day!

A very thoughtful reply, one which gets into the spirit of the question! 
Bravo! I really appreciate the work you put into this.

The question was "So, how would you go about solving the [environmental 
sustainability] problem?" if you had 105 million dollars per year, which 
is the UNEP's annual budget.

Let's consider your above reply as a first iteration. What I had in mind 
was what strategic approach would one take, at the managerial level. In 
a few clear sentences, what would be your strategy?

Now it's probably unfair for me to ask that question without first 
attempting to answer it myself. This would be the second iteration of 
answering the question. Here's a possible answer:


A workable strategy to solve the environmental sustainability problem 
would be:

1. Determine in depth what type of problem it is.

2. Develop a process that fits that type of problem.

3. Execute the process, while continuously improving it.


Naturally I have the System Improvement Process (SIP) in mind as a 
starting point for such a process. One thing you will notice about SIP 
is it does not fall into the "is there any" trap. It doesn't first try 
to generate a lot of solutions and then evaluate them. That usually 
fails on difficult problems. Instead, SIP first diagnoses the underlying 
cause of the problem. The is where the Root Cause concept, with its 
careful definition, is used. Once we know the root causes, then we 
identify the High Leverage Points (HLPs), that, when pushed, would 
resolve the root causes. Then, and only then, do we start to converge on 
a solution.

Because we now have deep knowledge of the problem's structure and we 
know exactly where to push, Solution Convergence is several orders of 
magnitude easier than the popular "are there any solutions that might 
work" approach. The reason is the original very large solution space is 
now relatively small. This is because once we know the HLPs, it's easy 
to think of many ways to push on them, several of which would probably 
work, because we don't have to push on HLPs very hard to make them work. 
By contrast, one has to push on Low Leverage Points (LLPs) so hard the 
effort is usually insufficient or unsustainable. This agrees with Jay 
Forrester's observation in Urban Dynamics on page 111, that a viable 
long term solution must overcome:

"The tendency of a [social] system to resist and counteract an applied 
force... Compensating counteraction can be disastrous if the applied 
programs are expensive. Only applied programs of intrinsic low cost are 
feasible."

You might wonder why I didn't suggest "Develop a process that fits the 
problem." If we can keep the process generic it's much more powerful. 
For example, notice how SIP can be used on hundreds of different 
sustainability related problems, as we drill down into the main problem. 
This eliminates the need to find or create a new process for each 
smaller problem. One standard process, with its many standard steps and 
terms, keeps a Tower of Babble from emerging from such a large, long 
work effort.

You might also wonder how this would lead to solution at only 105 
million dollars a year. Well, that constraint is part of the problem to 
solve. Based on my analysis so far, I personally think it's easy to find 
workable HLPs and solutions that would come in well under that budget. 
Remember, the budget does not have to include things like direct 
population or consumption reduction R&D or capital investment, as well 
as all sorts of other implementation costs, because a good solution will 
cause social agents to take that work and those expenses upon themselves.

Well, that's the second iteration. Your turn, Mike and others.

Jack
Posted by  Jack Harich <register at thwink.org>
posting date  Sun, 11 May 2008 17:47:15 -0400