There is a myth about the way in which science works. Scientists attempt to find out how natural systems work. From what little they can glean initially by unaided observation and by analogy with what they think are similar systems, they form hypotheses about the workings of the systems they are studying and put these hypotheses to the test. They do this by making predictions from their hypotheses and then checking these predictions via observations aided by scientific methods, which may include experiments, measurements, and so on. These methods often involve elaborate equipment, stringent controls and highly standardised procedures. Because of their sophistication they are thought to provide a transparent window on reality. They show us how things really are. Observations made, the predictions can be checked against the data and science moves a step forward: the hypothesis is confirmed or falsified and this general procedure repeated. In this way, it is thought, science moves ever closer, by successive approximations, to an understanding of how natural systems work.

Like many myths, this one contains some truth. But if research in the history and philosophy of science over the past half-century has shown anything of value, it has shown that the methods that scientists use to test their hypotheses are not transparent windows on the world.