Accident and Design: Contemporary Debates in Risk Management

“Four chapters good, two chapters bad” (with apologies to George Orwell) appears to have been orthodoxy's response to the 1992 [British] Royal Society publication, Risk: analysis, perception and mana ement. The “good” four chapters were those written by distinguished engineers, statisticians and natural scientists, which reflected a view of public risk management as properly the domain of science and engineering rather than of politics and economics. That is, both risk, and human behaviour in relation to risks, are objectively discoverable by orthodox canons of science, and the results, wherever possible expressed in numbers, are capable of being fed back into enlightened policy-making in the form of rational decision criteria applied by experts. The central problem for effective public risk management is to ensure that these decision criteria and the associated risk data are as accurate and scientifically well grounded as possible-including attempts to incorporate “the human factor”, provided that such an elusive element can be reduced to terms that are tractable for engineers.

Indeed, this approach is not simply an engineer's view, for there are social science approaches to risk that fit the same conception of risk management very closely, and hence can readily be absorbed into the paradigm. The most notable case is the well established and enormously successful psychometric paradigm, which aims to identify general features of individual cognitive approaches to risk; but Mary Douglas (1994) has argued that the developing “risk amplification” and “risk communication” approaches also fit with the dominant “enlightened engineering” approach to risk, because both essentially exclude politics from the analysis.

For this dominant approach, “problems” come in the form of “irrational” behaviour by politicians and bureaucracies, unaccountably declining public trust in scientific expertise, and “maverick” social scientists who do not see risk assessment and management as a politics-free zone. Indeed, the “two chapters bad” of the Royal Society document were precisely those concerned with risk perception and management. These chapters advanced the unorthodox ideas that there might be a political dimension in the way risks are socially construed and that the fundamental doctrines of risk management are in fact inherently plural, disputable and disputed. Significantly, the document was not issued as a report of the Society (unlike its 1983 predecessor), but rather as a report of a study group, with the contents of each of the chapters attributed to its author(s).

-xi-