modelling - an introduction

There are many reasons for drawing attention to modelling here. An obvious one is that in asking people to consider an alternative logic, it becomes easier if it is accepted that all logics are "just models" - models of causality. And models do not have to be true as such, simply useful in some pragmatic sense.

This is how we should view any scientific theory including biggies like Darwinian evolution, quantum mechanics or relativity. They are models of reality and can be judged by the extent to which they “work”.

Of course, what we mean by work then needs further examining. And here I find Robert Rosen’s modelling relation or Howard Pattee’s epistemic cut the most useful starting points. Or going a bit further back, the semiotic approach of philosopher CS Peirce on which pragmatism is based.

Anyway this is the epistemological reason for emphasising modelling. If all ideas about the world are “just models” then even our ideas about why anything happens are just models of causality. As long as we can form up our ideas with suitable crispness – define them axiomatically – then we can consider the merits of alternative logics.

But I have a second ontological reason for devoting a section to modelling. Epistemology is about how we can know things about the world. Ontology is then what we what we come to believe to be actually true – at least so far as the restrictions of modelling allow us to know anything at all.

The ontological reason is that I believe a theory of modelling is also going to be a theory of the mind. In its generalised way, it will be a theory about cognition, knowing, control, semiosis, or any other grade of "mindfulness" in a system.

So modelling theory is both a theory of human knowledge and also a more general theory of how any system is organised to have top-down “mindful” control. With pansemiosis, we can even talk about how universes have minds, or at least memories, habits and goals - the laws of physics as they are more usually known.

Peirce, Rosen and Pattee provide a solid foundation for talking about the modelling relation. But then of course what I want to add here is a dichotomistic twist. I want to bring out the fact that the modelling relation is dichotomistic (and hence a “1,2,3” journey from vagueness to crisp hierarchical order).

One of the key dichotomies is impressions~ideas. A mind develops by forming a set of generalised or global ideas that then frame, or rather constrain, the occurrence of localised impressions. A baby is born to a blank buzzing confusion. It then learns how to interpret the world. It develops the habits of perception that create a parade of crisp, brightly felt and meaning-imbued, mental impressions.

Another basic dichotomy is truth~control. Once we start talking about the purposes of modelling, examining what we mean when we say a model “works”, we will see that it dichotomises into two very different kinds of goals (which in turn are achieved by two very different kinds of mental architecture).

One goal is to know the truth of the world. To be objective in other words. To passively stand outside everything and see the whole of it. The other goal is to be able to control everything. This is the subjective pole of being. It is to be the “I” that can – God-like – make things happen. We could also say this is the contrast between the scientist as philosopher and the scientist as technologist. Or between organic and mechanical logic indeed.

The organic explains in holistic fashion. It is the more complete view. But holism then pays the penalty of often seeming overly complex - it delivers too much information, it requires too much memory. The mechanical approach to explanation can be almost comic-book reductionist. But it is at least maximally efficient. It is guaranteed to deliver control with the minimum of information, the least anount of memory.

Anyway, we will see that our understanding of modelling can be guided by a number of key  dichotomies such as impressions~ideas, control~truth, technology~philosophy, particulars~generals, mechanical~organic, subjective~objective, epistemology~ontology, knower~known, self~other, matter~mind, substance~form.

And note that here, as usual, local~global is our uber-dichotomy so all these dichotomies are listed with the local term first (like particular impressions or mechanistic control) and the global term second (such as general ideas or organic knowing).

<home> modelling> next