Why Model? | Santa Fe Institute

Any user of a map or model must realize that we do not understand a model, map, or reduction unless we understand and respect its limitations. If we don’t understand what the map does and doesn’t tell us, it can be useless or even dangerous.

In an age of global connectivity and existential threats to human life, computer simulations can help us model, interpret, and intervene in complex patterns of causality. We need to get better at making simulations, at playing with them, and at seeing through them to their underlying assumptions.

As Scott Page put it in his brilliant book The Model Thinker, for any complex phenomenon we need many models that can challenge each other, in this case including not just epidemiological models but also economic and social models.

A computer model of a process is, of course, a mathematical abstraction, not the thing itself, so a perennial danger is that the model doesn’t correctly represent the process—that critical but unappreciated details are left out of consideration—yielding misleading results.