Model to Navigate
There are at least two philosophies of how to model and what for. The one is to give a picture almost one-to-one of the matter modeled. For any complex phenomenon this is no philosophy to be recommended. It would just show the model becoming overwhelmed by the complexity it tries to model. ‘Complexity’ is the very concept to describe phenomena, systems, networks, what have you, as beyond modeling be it by means of mathematics, texts, stories, pictures, music, or dance.
The other philosophy is one of understanding models to be tools of navigation. It is not done externally, but internally, not outside-in but inside-out-and-back-in. This kind of model, be it by means of mathematics, texts, stories, pictures, music, or dance, should let you take decisions, even if via proxy or simulation, and lets you watch some of its consequences. It turns the modeler into an element of the phenomenon to be modeled, that is into a part of the world he or she is navigating in. It is a model to move within labyrinths, which are considered to stay labyrinths. You do not search for transparency, you look for possible next steps.
It seems to me that the second philosophy is the philosophy behind the idea of semantic webs as spelled out by Tim Berners-Lee and others as a way to somehow let the web self-organize via a programming code based on triples, most importantly the triple of subject, object, and predicate.
For social phenomena one might try to go for pragmatic webs instead of semantic webs. The basic idea here would not be grammar but action. And sociological theories of action tell you that there may be a way to model them by looking at their combination of self-reference, context, and paradox. Any action tries to repeat itself, depends on changing contexts to do so, and is made impossible and thus driven to look for variation by its own lack of foundation.
Thus, we have
as our metacode, which we can use to scan the routines and subroutines of social action in different spheres and different depth of focus as described by the rich corpus of sociological theory.
The metacode gives you the variables, not the values. One can get the values only by taking part, even if by simulation, aka ‘thinking’ or ‘theory’, with the phenomena one tries to control by controlling one’s own behavior toward them. This is operational research as understood by W. Ross Ashby many many years ago.