Models: Why Does Simplification Not Equal Distortion?

A model is not a miniature copy of reality. It is a structure designed for a particular question, deliberately preserving some relationships while omitting detail. A metro map leaves out street widths, building shapes and actual distances, yet highlights station order and interchange points. It may mislead someone looking for a café, but the same simplification is exactly what makes it useful for planning a journey.

A model is different from an example. An example presents one particular case; a model identifies relationships that can be used across many cases. Nor is every model a prediction. Some models are built to explain, compare or control without forecasting what will happen next.

We should therefore judge a model by more than how closely it resembles reality. We need to ask which question it serves, what it preserves, what it leaves out and where it fails. A good model does not contain the most detail. It makes the relevant structure clear enough to use while keeping its limits visible in our judgement.