Anatomy of doctors' beliefs

You go into hospital with symptoms of a potentially serious disease. They offer you a choice of two physicians: one always uses standard treatments, while the other is known to try all the newest ideas, however speculative. Which physician would you choose? It is easy to criticise doctors for being slow to change their beliefs and practices, but when you are on the receiving end you probably prefer conservatism in a doctor.

So, what conditions are - or should be - necessary to change a doctor's beliefs and practices? This raises fundamental questions about the nature of scientific knowledge and the processes of scientific discovery, the creation of hypotheses and the psychology of belief. Modern attempts to understand these processes began in the 17th century, with what is called Newtonian inductivism: scientists then believed that they induced laws of nature from experimental results. The next fundamental change was introduced in the 20th century by Karl Popper, who argued that the distinguishing feature of scientific knowledge is not that it can be proved, but that it can be disproved if counter-evidence is found. This theory provided a logic of scientific reasoning, but it leaves unexplained the creative process of scientific hypothesis formation.

The idea that science is influenced by the individual and collective psychology of scientists has recently been taken to an extreme by postmodernists such as Bruno Latour and Steve Woolgar. These writers treat science as a power game played by scientists, and they repudiate or ignore objective reality in science. Paul Thagard likens them to "an anthropologist who does fieldwork in an alien tribe without knowing the language".

In this book, Thagard tries to identify the processes involved in scientific hypothesis formation, discovery, experimentation and belief change. He does so by detailed analysis of one example. In the early 1980s,two physicians proposed that a big cause of peptic ulceration was a bacterial infection of the stomach lining. Their suggestion was far from the prevailing beliefs, and many dismissed it as ridiculous. Now it is almost universally accepted as true because of the weight of experimental evidence in its favour.

This book is remarkable for its clarity and its lack of doctrine. At each stage, Thagard outlines in plain terms precisely what he is trying to explain, and illustrates his explanation by reference mainly to the example of peptic ulcer, but also to Aids, scurvy, prion diseases and chronic fatigue syndrome. In successive chapters he considers the importance of serendipity, technology, scientific collaboration, analogy, the logic of inference and the reasons why a new scientific theory become accepted. Each chapter ends with a summary of the arguments presented, again in admirably clear language. I baulked only at the final chapter, in which Thagard proposes a model of science as distributed computing with scientists as nodes in the computing network: this seems unnecessary and contributes little to his thesis.

The book argues persuasively that the doctrinal theories of science are inadequate to account for the changes in beliefs about peptic ulceration in medicine that have occurred in the past 15 years. By allowing the importance of personality, ego and social pressures, as well as subscribing to a "medical realism" - ie the idea that there exists an objective reality in these diseases - Thagard sometimes runs the risk of seeming bland and non-committal. But it is precisely this even-handed and commonsense approach that allows him to give an accurate portrayal of what scientific advance is like. If this is what philosophers can do for science and medicine, we need more help from them.

Charles R. M. Bangham is professor of immunology, Imperial College School of Medicine.