On a Global Foresight thread Tom Abeles asked if scientific models, based on a reliable “bandwidth” for natural systems, might somehow have predictive properties. He noted:
The problem we have is exemplified by the poem, “The Theory that Jack built” which was published in an insightful set of “nonsense” poems called The Space Child’s Mother Goose: This is the theory that Jack built; this is the flaw, this is the constant covering the flaw; this is the “x” justifying the constant. . . und so weiter.
That’s a great insight and question. Part of the paradox seems to be in the word “like”. It is exactly models made “like” the ones for deterministic systems that don’t work well at all to imitate accumulatively self-organizing systems.
For new “unlike” models to still be useful, I think we need to no longer use models only as something we build to independently represent nature. I think we also need to use models for the opposite,
learning how to independently observe how individual systems in nature actually behave.
That produces predictive questions rather than answers, prompting where to look for answers regarding the individual behaviors of the well organized but non-deterministic world. Yes, it also sort of reverses the basic intent of science in a way, i.e. no longer treating the purpose of research as creating an artificial world made “like” the real one.
It has the potential to convert our large investment in deterministic science to shed more light on the kinds of new questions needed to explore the natural systems that display variable organization.
How observed “laws” of regular accumulative change can prompt a “change in laws”, with rules switching from one direction to another, is one of those.
Say you have a system in which the parts all need to grow to survive, and then find the system as a whole meeting constraints. How the parts will then react to the change in environment is not something that any model of their prior cooperative growth before meeting limits can offer.
Models don’t show the limits of the system they are models of,
or it’s future states, just it’s past state.
A way to use that flaw is to consider systems as locally adapting to their environment as it changes, reacting on their own. That would treat them as hypothetically changing within the deterministic limits offer a “bandwidth” for their individual behaviors.
As they approach or cross those limits in a progressive way, it prompts useful new questions. As often:
knowing what new questions are coming
will expose a few that can be answered.