Many scientific hypotheses are nowadays granted with a social, political or commercial function. They are prematurely released from the laboratory, without full support from empirical evidence, but with a specific task: to warn the world about dangerous situations or evolutions, or to inform it about promising trends and capacities. And, whether in the area of environmental protection, economics, health or technology assessment, in many cases, they are produced as ‘if-then’ hypotheses upon explicit request from politics, civil society or the market. In the absence of evidence that would facilitate straightforward judgement, consensus and consequent action, these hypotheses have themselves become the ‘end products’ of science, and society has no other choice than to deal with them…
The argument developed in this paper is that the more fundamental problem with science that aims to advise policy today is not the problem of strategic manipulation of scientific advice by politics, civil society or the market (or by science itself) per se, but rather the challenge of dealing with the lack of evidence in situations where politics, civil society or the market ‘need’ that evidence to (eventually urgently) inform, criticise or justify specific actions or practices. All concerned actors, in these cases, are left with nothing more than underdetermined scientific hypotheses. And in the absence of evidence that would facilitate straightforward judgement, consensus and consequent action, these hypotheses have themselves become the ‘end products’ of science and they start to lead an own ‘political life’ themselves. As a consequence, society in general, and science, politics, civil society and the market in particular, have no other choice than to somehow deal with them in a responsible way.
Based on a number of cases, the paper characterises the problem sketched above, discusses a number of relevant academic views on what science is, can be and should be today, and consequently elaborates on what that ‘new responsibility’ would need to be.
> > >