OUR CONCLUSION, IN PART I,* WAS THAT THE ABANDONMENT OF THE expert/lay dichotomy as the basis for understanding risk perception, whilst essential, is not going to be easy. We argued that:
1) Objectivism (the idea that we can clearly distinguish between what the risks really are and what people variously and erroneously believe them to be) has to give way to constructivism (the idea that risk is inherently subjective: something that we project onto whatever it is that is ‘out there’).
2) To impose a single definition of what the problem is, which is what so much of policy analysis and science-for-public-policy does, is to exclude all those who happen not to share that particular way of framing things. Since people are unlikely to support a policy that is aimed at solving what they do not see to be the problem, approaches that insist on singularity (and on single metrics — cost: benefit analysis, for instance, probabilistic risk assessment, qualityadjusted life years and so on) will inevitably be low on consent, surprise-prone, unref lexive, brittle and undemocratic.