rom[Bill Williams 2 June 2004 9:20 PM CST]
from Mary Powers 2004.06.02
[From Bill Powers (2004.06.01.0757 MDT)]
As I am unable to communicate my ideas about how to start modeling
Economic systems or anything else (the
defenses are too much for me to overcome)...
* * *
Yes, it is a control thing. The strength of the resistance depends on the
importance of what is being defended. I'd say principles at the very least,
more likely systems concepts.
It is Mary Powers argument that recent disputes have as their source a "resistance" to Bill Powers conceptions concerning modeling. I think it fair to characterize Mary's statement as a claim that there is no merit at all to recent Posts by Martin Taylor, Bruce Nevin, or myself that take issue with Bill Powers in regard to some fundamental issues regarding social theory.
However why not ignore these questions for the time being. There is I think an issue underlying these differences that might be in some sense prior to arguments between Bill Powers on one side and Martin, Bruce Nevin, and myself making up a more or less the other side.
Consider what Bill Powers says about his failure to communicate his ideas about modeling. First a question: Where would one go to familiarize oneself with Bill Powers' conception of modeling?
The only place I know of to go for Bill Powers' views that come close to bearing upon fundamental issues regarding modeling is the Appendix to _Behavior: the Control of Perception_.
I could very well be missing some information, but I think that having been around when Greg Williams was editing the publication of Bill's papers for publication in the Living Control systems sequence of books that I would have learned of any statement by Bill Powers concerning modeling. But, I am willing to be corrected if there is a better statement of fundamental issues than in the appendix to _B:CP_.
Now, I am sure that most of the long term GSGnet subscribers have read and carefully studied the appendix to _B:CP_. When I first encounter control theory it was by way of William Glasser's _Stations of the Mind_. I found Glasser's book fascinating, I picked up a copy in Boulder on the way to the airport and read it on a plane going to St. Louis. Arriving at St. Louis I immediately ordered a copy of _B:CP_. When I got a copy of _B:CP_ one of the first things I did was look at the appendix. And, it was obvious to me from the appendix that Bill Powers was going at his analysis in a way that was very different than any psychologists that I had ever encountered. And, I had by that time looked at a lot of psychology. And, I had looked without finding any theoretical work in psychology that would make any fundamental contribution to reconstructing economic theory. But, I could see right away based on the appendix to _B:CP_ that this should provide the foundation for a new conception of economic theory.
However, if you haven't recently read the appendix to _B:CP_ take a look at it. And, then ask yourself, to whom does this passage communicate? How does it communicate? And, then go on to ask, What does it communicate?
For me there are two crucial sentences in the appendix. One occurs on page 281 which says, in part,
"...time itself is left out of the sequential-state analysis."
The other crucial sentence is located on page 273 and it says,
"For more advanced information see any text listed under "Servomechanisms" or
control Systems."
While the sentence is entirely correct in the sense that undoubtedly texts on servo control systems contain more advanced information, this referral has something of the "Elephant Soup" recipe joke, that begins by telling you to catch an elephant.
But, I followed the instructions, more or less, by learning about Op-Amps. Op-Amps are a kind of amplifier. You can buy basic versions say a 741 for about a quarter. With an Op-Amp and some other components you can build feedback circuits and learn with real devices how a control system works. And, after a while, if you are curious you can begin to think about how starting with transistors and other basic electronic devices you could build an Op-Amp from scratch, or at least from much more basic devices.
And, if you know even a little bit about this stuff there are people around, engineers even who have had the servo-mechanics course-- and even written a thesis on servo mechanics but don't really understand servo-mechanics that need your help. Actually doing the design work, as I did, for a servo system can be an education all in itself. Would you suppose if you use the same hydraulic pressure line as a sensor and also as an output that this might cause difficulty? Now that I think about it, I had the thought-- that can't possibly work. It took me a moment to understand how. With the right filtering the time domains (if this is the proper term) can be so different in frequency that it works just fine.
It takes some digging and it takes quite a lot of study but you can get a sense, or at least some sense, for what is going on. And, while designing control system for a Soya Bean harvester isn't rocket science, it was a job that the Engineer who had actually taken the Servo-mechanics course, and written a thesis on a servomechanisms problem couldn't do. The only reason I had a job, was because he couldn't do it. It may not be that tough a problem, but think about what is involved if the controlled variable the sensor pressure is read from the same hydraulic line that is used to deliver power to the output actuator. And, the output is delivered as an on/off pulse. I thought for a while that given the problems involved that someone was quickly going to figure out that it wasn't likely that I wasn't going to solve them. This wasn't the only problem involved. For some reason the Engineer would not give me a copy of the schematic for the whetstone bridge type pressure sensor. Eventually, after some severe headaches, I got the system working. And, despite some rather peculiar features it worked quite well.
I don't think this is the right time to go into another control theory issue, at least not in detail, where I ended up as the technical expert ( sans credentials ) for a national problem that a certain make airliner had with its electrical system. However, I do wish to at least mention this encounter.
I never found out who designed the electrical system for the DH 112 Heron, but it was obvious that who ever it was they didn't understand control theory, and neither did the guys who certified the system. When I finally found an engineer who really did understand control theory and explained how the system was set up he started laughing. I suggested some changes in the system, but the company involved didn't wish to take responsibility. I then devised a procedure that kept the system adjusted so that it had a chance of functioning. The pilots of that piece of junk, however, didn't think it was all that funny. And, the passengers who died as a result of an inopportune failure the electrical system probably felt that they had been inconvenienced. By-the-way, the pilots involved were charged with responsibility for the accident because they took off with a system they knew was defective. But, that is another story.
The point I am attempting make is that while I do not have a formal background in control theory I have had some experience with getting
control systems to work.
I would also point to my having developed some modestly complex programs,
involving control theory- the Lattice program is an example. Rick as a
PCT theorist has expressed doubts that the program actually contains
control loops. This difference of opinion doesn't necessarily mean that on the whole I understand control theory better than Rick. Or that I am a better modeler, on the whole, than Rick. However, if it turns out, as it will, that there actually are control loops in the program, then this may indicate that I understand some aspects of control theory that Rick does not understand.
Now, the Lattice program, particularly the contrast between the A and the B versions illustrates a principle that is presented in the appendix to _B:CP_ .
Mainly that a delay in a feedback loop can contribute to the instability of the loop. This at least is my understanding of the disturbance that grows as it travels across the Lattice.
Now to return to Mary's argument that there has been "resistance" to Bill Powers conception of modeling.
Where in an organized text has Bill Powers presented his conception of modeling? In terms of a systematic presentation, I don't see that there
is anything that I or anyone else could "resist." I regard to the actual
process of modeling, we haven't as yet, been confronted by anything to
which we could resist.
There has, however, been on Bill Powers part a tendency to express what
I regard as arbitrary opinions concerning various economic issues-- such
as the claim that "It won't cost anything to go to Mars." And, opinions
too about Keynes that have no connection at all that I can see to control
theory. Nor do I see that "resistance" to modeling has anything to do with
the objections by Martin Taylor, Bruce Nevin, or myself to Bill Powers'
opinions concerning social theory.
At least from my perspective Mary is making a mistake to characterize the
situation in terms an irrational "resistance" to truths concerning modeling.
First, where have these principles been expressed? And, second, I would
argue that recent disputes do not have as their source a "resistance" to
any systematic presentation of the principles of modeling-- because no such
presentation is, to my knowledge, available.
Bill Williams
Not a peep though from anyone.