Models

(This is not a post about the fashion business. Nor, despite deriving inspiration from discussions of an economic character, is it directed at such discussions exclusively).

We use models all the time. We have no choice: it is the principal way we humans have of putting order on, making sense of, the world of experience. No-one decides to do it, and it is pretty much impossible to decide not to do it. In the current jargon, we are "hard-wired" to be modellers.

Some of the models that we use were designed by others, but when a perfectly suitable one is not to hand, we will design our own. This is very rarely done entirely "from scratch": we usually adapt models we receive from others. (In this latter way, we can all be slaves to the ideas of some long-dead philosopher in Keynes' memorable phrase, especially if we fail to notice).

Models are not always so-called. Another word for the same thing, albeit not usually as formalised, is "narrative", or even less formally as in Tyler Cowen's talk in this TED presentation, which I strongly commend to you, as "stories". "Framing" is a cognate activity. "Paradigm" is yet another word for a model of this type.

Less benign or neutral words are "stereotyping" and "stigmatising", "labelling" - all of these are also consequences of a modelling process.

Whatever name is used, the process is similar. I am going to call it modelling.

The point that I want to make about modelling is that it is both inevitable, and yet inevitably is just as fallible as all human cognitive activity is. Only the simplest models remain useful for long; one that lasts even one normal human life-span is unusual. It is prudent to be sceptical of all models, not in a negative or destructive way (because all models have some useful function for someone at some time) but to avoid falling into error through over-attachment to them.

Why ? Because a model incorporates beliefs about how things are. Many of these beliefs will turn out to be incorrect. (In some formal models, the truth of the beliefs is explicitly a mere assumption.In less formal models, mere assumptions are unfortunately rife, and, worse, unacknowledged. Worse again, this is fairly common even in formal models).

Dangerously, a model may continue to work well for a long period even though it incorporates assumptions which no longer correspond as closely to reality as they may once have done.

Of course, this is a counsel of perfection: no-one can avoid this trap entirely. Even Socrates, who, as far I know, first formulated the radical scepticism which I am advocating, was, after a fashion, killed by his refusal to abandon a fatally flawed model.

So-called "scientific method" is a powerful tool in the service of this Socratic endeavour. It is not fashionable or popular in certain quarters to say so, but science, as we now understand the word, is not primarily a constructive activity: most of the time of scientists is spent discovering flaws in the models, sometimes their own, used to describe reality.