The Post-Normal Perspective, September 2003
Notes for the lecture:
Pluralistic uncertainty management: social scientific basis.
EU Advanced Summer Course in Integrated Assessment
ICIS, University of Maastricht, September 2003.
1. The context.
We are now going through a revolution in epistemology: science does not deliver certainty. The faith established by Descartes and Galileo is overthrown. In policy issues, the facts are frequently inconclusive; they are subordinate to problem-choice, which itself is influenced by issue-framings and their defining value-commitments. In many science-policy issues, our ignorance is more important than our knowledge. Radical uncertainty rules. After centuries of triumphalism, the rule of dogmatic science is at an end.
While the legitimacy of actions of the modern state depends strongly on public welfare, the authority for its policies depends in the last resort not on divine sanction or birth or wealth (as in the past), but on science. When official science is revealed as incompetent or corrupt, there is a recognised 'crisis of trust'. When people ask, "Why should we begin to trust you now?" after scandals like BSE, the crisis of governance is real. Also, when the public discovers the extent to which the maintenance of power depends on the social construction of ignorance, in which the taming of science plays an important part, the foundations of acquiescence are further eroded.
Among enlightened authorities including the European Commission itself, it is recognised that a 'democracy of expertise' must be established. Otherwise, 'the consent of the governed' will not be secured, and the workings of the modern democratic state will be threatened. Paradoxically, the Non-Governmental Organisations are being drawn into the processes of governance. To some extent they are being co-opted, but they also retain their identity and their partly antagonistic function.
The new organising theme for the conduct of science in the policy domain is 'debate'. This is a recognition that policy issues, even those including science, are not merely uncertain but also complex. That is, there is a plurality of legitimate perspectives, not reducible to a single dominant 'correct' view. The task for science is not to achieve a truth to which all must subscribe, but to establish a basis for negotiation in good faith.
All these developments are comprehended by the theory of post-normal science. This applies when, typically, 'facts are uncertain, values in dispute, stakes high and decisions urgent'. There is a need for an 'extended peer community' for assuring the quality of both process and product. In their creative engagement, they will deploy their 'extended facts', which can include anecdotal evidence, confidential information, local knowledge and ethical commitments.
Under these conditions, we move from the dream of conclusive scientific demonstration to the ideal of dialogue for reconciling real antagonisms. The truncated self-awareness of scientific expertise, necessary for its reductionist strategy, is replaced by an enhanced self-awareness of participants, which includes their own uncertainties and commitments.
This awareness includes the recognition of the inescapable value-loading of every inference, statistical or scientific. The choice of confidence-limit in statistical tests embodies the value-laden decision between the errors of excess selectivity and of excess sensitivity. When, as is now typical, scientific evidence is inconclusive, there is a policy-driven choice between principles of inference. One principle is that absence of conclusive evidence of harm is equivalent to conclusive evidence of absence of harm. The other is that absence of convincing evidence of harm is equivalent to suggestive evidence of official concealment of harm. Methodology is now politicised, and is itself part of the post-normal processes of debate.
This debate now has a characteristic asymmetric structure. The reductionist sciences produce innovation, whose proponents attempt to deny danger. The wholistic, post-normal sciences call for 'precaution'. On the globalised corporate side, ignorance is ignored or discounted; on the 'safety' side, it is emphasised as critical. The nominally independent institutions of the traditional research community are caught in the middle. If they succumb to the pressures to support innovation, they are at a serious risk of losing the public's trust; but if they resist those pressures, they can be penalised in many ways.
The sciences of safety develop their own methodologies, which can be seen in their characteristic leading questions. In the search for knowledge, they are 'what/how?'; and in the development of devices, they are 'how/why?'. In the post-normal sciences of safety, the leading questions are 'what-if?'. Such new questions are calling into being the corresponding new techniques of inquiry and with them the appropriate social institutions and practices. They are the natural questions for members of the extended peer community, just as they are strange and unnatural for those trained in the myopic puzzle-solving of 'normal science'. On the social side, 'community research' is a vigorous new growth.
There is now a great new organising theme for this sort of science: safety. The traditional goals of Western science, knowledge and power, are now compromised. In dialectical fashion, their successes in improving the living conditions of at least the world's rich, have produced the need for a new ideal. For their successes in producing safety, locally and in the short run, have created new dangers for the whole planet and in the long run. The attempted reduction of 'safety' to 'acceptable risk' will prove as unsuccessful as the nuclear power industry for which it was developed. By contrast, 'safety' is a complex, post-normal concept, which is at once pragmatic, recursive and ethical.
3. Tools and techniques.
On the social side, there is a growing family of methodologies for structuring dialogue on the post-normal issues. Some have grown out of initiatives like the European Community projects ULYSSES and VISIONS. There will be many examples discussed in this course. In all of them, there is an attempt at integrating, indeed at reconciling, the very different sorts of knowledge characteristic of the scientific approach, including IT techniques, and that of citizens in an extended peer community.
On the more narrow front of the technical management of uncertainty, there is the NUSAP family of methods of assessment of the qualitative aspects of quantitative information. This has now been developed into a very powerful tool by the collaboration of Jeroen van der Sluijs at Utrecht University with colleagues at RIVM, notably Arthur Petersen. Through a set of checklists, this provides practitioners in any field with the means to express and communicate their personal, craftsmen's awareness of all the qualitative aspects of their quantitative information. It also provides the users of such information with a perspective on its strengths, limitations and weaknesses. In that way, it contributes both to its more effective utilisation and to strategies for its improvement.
The technical achievement, in the creation of such tools for the enhancement of understanding, is considerable. The philosophical contribution of this approach is even more important. It shows that the demise of the long-standing dogmatism of science need not lead to a post-modern anarchy of nihilism or relativism about facts, values and reality. We can grant that scientific knowledge, like any other image of reality, simultaneously reveals, distorts and conceals. The great philosophical challenge of our time is to comprehend these apparently contradictory, but actually complementary, aspects of knowledge. With a holistic, systems conception of knowledge itself, it is possible to begin the reconstruction of our philosophy of scientific knowledge, with its varied dimensions of knowledge, power and experience.
That, for me, is the main lesson of the plurality of uncertainty management, as a contribution to this summer course in integrated assessment.