Commentary: Facts, opinions and affaires du coeur

Abstract

It is now over 30 years since Michael Marmot began his comment article with the provocative quotation ‘. . . these are the opinions on which I base my facts . . .’. Here Marmot identifies a central tension found in many branches of the sciences: namely, how is it possible to evaluate and choose between the competing theories or hypotheses that often seek to explain the same phenomena? A concern that is more germane than ever: he asks if, despite all their scientific trappings, the decisions taken by scientists are not simply based on the irrational and the intuitive. And, as Marmot points out, this is not merely an academic argument, it has significance for understanding the causes of disease that are of deep importance to a wider public. He uses a case study of coronary heart disease (CHD) to suggest an answer to the thorny question of how to choose between competing theories. Four accounts, taken from the philosophy of science, were chosen to shed light on this issue: the inductive view; the Popperian view of science; the theories of Lakatos; and Kuhn’s account of revolutionary science. The inductive view, that once sufficient ‘facts’ are gathered a ‘correct’ theory will emerge, was dismissed as essentially hollow—theories do not emerge fully formed from the dry dust of facts and the ‘history of science commonly gets reconstructed from an inductive viewpoint’. The Popperian view that theories are tested by attempts to falsify them, while more attractive is equally doomed because it is always possible to save a theory by the presentation of an ‘auxiliary hypothesis’. Marmot finds the remaining two accounts of science to be more productive. The Kuhnian view is that in periods of ‘normal science’ the work of most scientists consists of tinkering and problem solving—in other words adding footnotes to an established ‘paradigm’. Periods of revolutionary science, in contrast, are marked by the accumulation of anomalies and the breakdown of an existing paradigm and the emergence of a new one. This new scientific knowledge would, according to Kuhn, be ‘incommensurate’ with the old paradigm—like all revolutions those on the losing side would not even be able to comprehend what had overtaken them. Whereas this view may throw light on some developments in scientific knowledge, it is doubtful if it has much to offer in terms of understanding debates concerning CHD—as these, from a Kuhnian point of view, largely represent problem solving within the ‘normal science’ of epidemiology. Indeed, periods of revolutionary science probably occur rather infrequently (one such example might be the increasing influence of modern epidemiology on medicine and public health policy in the 20th century). Marmot finds that the theories of Lakatos provide the most compelling aid to judging between competing research findings. In this view, an account can be retained as long as it remains progressive—a theory that encounters anomalies can be thought of as progressive as long as these point towards new empirical observations. Thus, using Marmot’s example, if one believes a causal factor in CHD is fat intake and one discovers a group with a high fat intake but low levels of CHD, it might indicate that not all types of fat cause disease. In other words, this anomaly points towards a new empirical observation that might allow the theory to remain progressive. This theory, Marmot argues, is a valuable insight from the philosophy of science because it provides some practical help in choosing between competing explanations. One of the central arguments running through the Marmot article is that scientists in general and epidemiologists in particular, can learn valuable lessons from studying the history and philosophy of science. In the period since his article appeared, many anthropologists and social scientists have increasingly turned their attentions towards the natural sciences and scientific practices and given birth to a new sub-discipline—science and technology studies. What insights might be now be added to complement Marmot’s call to take ‘philosophic speculations’ seriously? One of the first is that the nature of knowledge production and consumption has increased quantitatively and altered qualitatively since the mid-1970s— for example, the World Wide Web did not exist in any meaningful way before the 1990s. It is now no longer Department of Sociology, The Open University, Milton Keynes, UK. E-mail: s.d.carter@open.ac.uk Published by Oxford University Press on behalf of the International Epidemiological Association