[From Bruce Nevin (2017012020.15:45 ET)]
I’ve started reading again in the correspondence between Bill Powers and Phil Runkel, which Dag in his quietly heroic way has published as Dialogue Concerning the Two Chief Approaches to a Science of Life. Highly recommended reading.
The following is from Bill Powers’ letter to Phil Runkel on 4 December 1987. This letter is an extended review of the manuscript of Phil’s book Casting Nets and Testing Specimens. The paragraphs below begin his discussion of Phil’s Chapter on specimens and control theory.
I think it is essential to follow the course that Marken set. First we must establish control as a phenomenon. This is not a theoretical matter. We have to show that organisms actually do stabilize external variables of all degrees of complexity against disturbances, maintaining them recognizably near reference conditions that we can identify experimentally. We have to show that the relationship among controlled variables, disturbances, and actions is a real relationship, a directly observable fact of nature. No theory is needed in order to do this. The fact is that organisms do behave in this way. This observation has nothing to [do with?] how they could behave this way and still be physical systems.
This is precisely where psychology went astray. Psychologists observed this phenomenon, although they didn’t observe it very competently, and chose to disbelieve what they saw because it went against principles they had decided to treat as holy and superior to the data. Essentially all the contortions of psychological theories and philosophies of science have been generated exactly to explain how it is that behavior can appear purposive yet not actually be purposive. I think the miserable record of the life sciences hinges on this fateful choice to ignore the data.
In any case we control theorists have to establish the reality of the observations first. Then we can raise the question of finding a theory that makes sense of them. Fortunately, this theory exists in mature form …
So first we have the phenomenon of control. Then we have the theory of control systems. Then we have the model buil[t] on that theory to account for more and more of behavior…
For those who have contested the point, Bill here unequivocally asserts that controlled variables are “external” in the public environment where they may be observed and measured, along with disturbances and actions and the relationship among these three.
And for those who have caviled about the term ‘circular causation’, that is in Bill’s discussion of Phil’s Chapter 9, which in Phil’s book as published is titled “Specimens and Circular Causation”.
···
/Bruce