THE ROLE OF MODELS IN PHYSICS*

MOST scientists today are not very satisfied with the account of scientific method given by the philosophers. It is certainly true to say that the presentations found in books on the philosophy of science still show the traces of the Bacon-Mill tradition. But even where the misconceptions of epistemology have been left behind, there remains some basis for dissatisfaction : for the scientific method talked about by the philosophers seems strangely remote from what scientists believe to be the method they follow. I must hasten to add that it is not necessary to assume that the scientists are always right in their criticism. The scientist can often do his work widiout an explicit appeal to general concepts and to logic ; training and skill guide him to his result. Experimentalists therefore tend to confuse the practical procedures employed in obtaining results with the theoretical methods designed to explain these results. But, I think, most will agree that all present methodologies leave a wide gap. To fill it we must keep as close as possible to the expressions actually used by scientists. We must take as an example a modern theory and discuss actual laws, instead of illustrating scientific method by means of old-fashioned and very simplified examples. Indeed, what is often exhibited as an instance of a scientific law, e.g. the sentence ' All swans are white', can hardly be recognised by anyone as even remotely resembling such a law. Surely we never have exceptions to the law so conveniently at our disposal (like black swans) when we consider, for example, Newton's laws of motion. The use of such biased and over-simplified examples has seriously affected our understanding of scientific method. Similarly, the theories usually offered as specimens represent not even a small part of what, at least for physicists, is the simplest theory, i.e. Newtonian mechanics. However, I must again come to the defence of the philosophers, at least to some extent. For scientific theories, even Newton's mechanics, are very complex ; and more modern theories are still