Abstract

In the first instance this paper defends an instrumentalist account of the laws of science. This is opposed to what I call the ‘vending machine’ view: pop some initial conditions in—say, about the past history of your system and its setting, let the laws run and out comes a new result—say a description of its current state or a prediction about its future. We all know it isn’t like this. Establishing principles and laws in science is tough work. And so too is putting them to use. In takes ingenuity and practice, know-how and luck to use our laws to learn new facts. So scientists, I claim, are artful modellers, and laws are instruments science has learned how to use to build good models for prediction, explanation and manipulation.

What then becomes of the laws of nature that the laws of science were supposed to ape? These laws have long been pictured to be like detailed bureaucratic principles that dictate what should happen in each and every circumstance, down to the last detail, no judgment, no variation, no fun. If we build our image of what nature is like from what we do in our most successful scientific endeavours to describe and predict—as I think empiricism dictates we should—this picture of nature the fastidious and totally thorough bureaucrat is hard to believe. More natural to science as it is actually practised is rather that nature like us is an artful modeller. The facts do not after all accord with some hidden and thorough set of bureaucratic principles. They are more varied and relate in far more complex ways than any such principles can describe. The best possible way to recoup the facts may, even in God’s heaven, be by artful modelling.

This talk will try to explain these two ideas and show why we should take them seriously.