[Date Prev][Date Next][Thread Prev][Thread Next]   [Date Index] [Thread Index] [Author Index

Re: Metabolism/Repair and Modeling Relation



John M,
 
I think your confusion is understandable, especially when M,R and M-R and MR etc are thrown around in discussions about Rosennean Complexity Theory, relying on the context of the sentence to clarify meaning... except that in my father's work the context could seem to point either way. But I think I can clear up the confusion here. Let me know if this helps:
 
The model my father created to illustrate the bare minimum necessary behaviors of a system in order to be a living organism is his (M,R)-System diagram. In this diagram, M stands for Metabolism and R stands for Repair. The diagram (and name) itself is an example of a modeling relation: formalism (diagram) to natural system (organism).
 
There is another diagram my father created, of the modeling relation itself. In it, the natural system has "causal loops" of entailment and the model has "impredicative loops" of entailment. There's an arrow going from the natural system to the model that is labeled "encoding" and another arrow going from the model back to the natural system, labeled "decoding". Those arrows embody the relation between a model and its corresponding natural system: a modeling relation. (Perhaps it's worth pointing out that any model is only as good as the encoding/decoding that relates it to the natural system it models. It obviously behooves us to get it as right as possible.)
 
You can also have a modeling relation between two models (for example; between a formalism on paper and a laboratory rat-- which is standing in as a model of a human being-- in medical research). This is where it gets very dangerous: The more times removed from the original natural system being modeled, the more inaccuracies will become part of the results predicted by the models.
 
The reason there is so much written in all my father's books about the subject of the modeling relation (metaphor, similarity, optimality, computability, etc) is because modeling is one of the main tools of science; BUT if it is done poorly then what we learn from it is of questionable value. Worse, if the model becomes the focus and the original system is forgotten.... then we learn about the model, not the  natural system it models. But the decisions we make based on the model are going to be applied to the natural system. Doing this generates "side effects". The disconnect between the model, then, and the natural system is a disconnect in the "modeling relation".
 
Judith